A Model-Recommended Win Total for Your Consideration

Fun fact: I have never bet on a season win total before. I admittedly always thought they were a bit pointless. Even if I thought a team’s win total over/under had value, why would I tie up my money for an entire season when I could just leverage my disagreement from game to game? It’s a legitimate critique that I wouldn’t fault any bettor for expressing. But given my relatively recent foray into predictive sports modeling, I’ve had the opportunity to evaluate team win totals in a more concrete way.

Seeing in hindsight the profit I could have accumulated with my NFL model by taking action on large model disagreements gives me the assurance that doing that exact thing with the MLB model I’ve built for the upcoming 2019 season should have positive expected value. And I figured with you guys waiting for my impending announcement of what my plans will be with the MLB model, I thought I’d share some of that information with you and give you something to chew on.

As a Cubs fan, seeing that the Cubs were the most overvalued team in the team totals market hurt quite a bit, but it’s really not that hard to see how and why this team would be overvalued. In terms of offseason moves, the Cubs did absolutely nothing of substance whereas the rest of the NL Central did their best to ensure this year’s divisional race is a bloodbath. The Cardinals added Paul Goldschmidt and Andrew Miller; the Brewers added Yasmani Grandal and Mike Moustakas; and the Reds added Tanner Roark, Yasiel Puig, Sonny Gray, and Matt Kemp (while getting rid of Homer Bailey).

On top of that, they have a bunch of talent that is either already regressing or are prime regression candidates (Jon Lester, Cole Hamels, etc.). Javier Baez in particular is due for regression in the biggest of ways this season. Last year he batted .290 for 34 HRs and led the league with 111 RBIs, which was a massive improvement over his 2017 campaign in which he batted .273 for 23 HRs and 75 RBIs. Granted Baez did have 27% more plate appearances in 2018, but the bump in his counting stats production vastly outpaced the rate of his opportunity increase.

That production increase came despite maintaining a similar and abnormally high BABIP (.345 in 2017, .347 in 2018), and can be best explained by his increase in power (.207 ISO in 2017, .265 in 2018) which then led to the massive jump in his slugging percentage (.480 in 2017, .554 in 2018) and HRs. Despite those spikes, his BaseRuns only jumped from 3.8 to 3.9 across those two seasons and it’d be much more likely for Baez to have a 2019 that looks more like 2017 than 2018.

And while the Cubs have a ton of talent trending downwards, the rest of the NL Central has budding talent. The Cardinals have Marcell Ozuna, Jack Flaherty, and Alex Reyes; the Brewers have Ben Gamel and Keston Hiura; and the Reds have Luis Castillo, Nick Senzel, and a Sonny Gray without a non-destructive pitching coach. Considering 35% of the Cubs’ games will be against the Cardinals, Brewers, and Reds, they have a tough task to get wins as-is.

Another 20% of their games will be against the NL East, which is full of playoff contenders (ATL, PHI, NYM, WAS). A look at the supporting data doesn’t make their case any better, as the model projects this team to only be a top ten team in starting pitching while ranking league-average in run production, tenth-worst in relief pitching, and near the bottom third in fielding. That doesn’t sound like a team that should be tied for the seventh-hightest win total, and the model agrees. Take the Chicago Cubs to go under 88.5 wins.

Kyle Freeland’s Very Weird (and Profitable) Home/Away Splits

In last week’s write-up we took a look at the offensive side of baseball and the continued evolution of batting metrics, and for this week’s write-up I originally was going to do a similarly-structured write-up with the pitching side of sabermetrics. However, after giving it a second thought I thought I’d switch it up and use this topic as an opportunity to just demonstrate how modern day pitching metrics can be utilized when evaluating a pitcher and his development while also using it as an excuse to talk about my favorite pitcher in all of baseball, Kyle Freeland.

Kyle Freeland is an anomaly that I could probably write an entire book about, and he has only played two full seasons of major league ball. In 2017, Kyle Freeland put together an above average season for a rookie starting pitcher. In the following season, Kyle Freeland took a monstrous leap and became one of the best pitchers in the National League, finishing fourth in NL Cy Young voting. He markedly improved in basically every statistical category fathomable as you can see below.

NOTE: The “minus” metrics for pitching are similar to the “plus” metrics for batting which we covered in the previous write-up, in that they are park and league adjusted and are scaled in a way that 100 is league average. However, the inverse is true with the “minus” stats in that each point below 100 represents the percent better a pitcher is than the league average, and each point above 100 represents the percent worse a pitcher is than the league average.

But it isn’t Kyle Freeland’s incredible development as a pitcher that is what is most fascinating about him. What makes Kyle Freeland so fascinating is that despite having the misfortune of pitching in the extremely hitter-friendly Coors Field for half of his games, his home/away splits show that he is actually a much better pitcher at Coors.

The result of these astonishing splits is my favorite betting “trend” ever: In Kyle Freeland’s starts in his first two seasons:

  • The under went 41-16 for +23.4 units (37% ROI)
  • The under in home games went 25-5 for +19.5 units (59% ROI)
  • The F5 under in home games went 25-4-1 for +20.6 units (62% ROI)

Like most betting “trends”, the Kyle Freeland under “trend” is not some magical force that leads to guaranteed profits over time for the rest of eternity. Instead, it is merely a reflection of a previously uncorrected and/or inefficient pricing pattern or strategy in the betting market. To help illustrate this, lets take a look at some of the underlying elements that may have led to this incredibly profitable trend with Kyle Freeland’s starts.

The first thing that makes Kyle Freeland’s success in his first two seasons and his success at hitter-friendly Coors Field during that time so astounding is a 25 year old performance trend of pitchers drafted by the Rockies. As you can see in the below, not only are Rockies-drafted pitchers dead last in WARP accrued with their drafted team, they are the only team with a negative figure in that regard. The forward-projecting sentiment that comes as a result of this historical performance can explain a part of the undervaluing and mispricing of Kyle Freeland’s starts, especially the starts from his 2017 rookie season.

Part of Freeland’s comfort at Coors can be attributed to the fact that the ace was born and raised in Denver, allowing him to be acclimated essentially since birth to the effects of the altitude. As a result, Kyle Freeland experiences a Freaky Friday-like home/road phenomenon in which the altitude and conditions of Coors is his “comfort zone” while pitching elsewhere is what pitching at Coors is like for the the rest of the league. Nevertheless, whatever adjustments betting markets made following the 2017 season to better capture Freeland’s comfort at home was never going to be enough to capture his development as a pitcher heading into and during the 2018 season.

The biggest change for Kyle Freeland in 2018 was his pitch selection. A large reason for this shift is directly tied to the development of sabermetrics and its ability over the years to identify which pitches work better and worse at Coors due to the conditions. In summary for the unaware, curveballs and sinkers have a measurably sharp drop-off in performance whereas sliders and cutters are proposed as better alternatives. As you can see in the above, Freeland opted to cut his sinker usage by over half in 2018 and reallocated that volume to the rest of his more Coors-friendly arsenal. The change was certainly deliberate as Freeland himself noted, “Last year we discovered after the first half that guys were looking for sinkers down and away, because they knew I would be throwing them, and I started getting hurt throwing those pitches”. Freeland also worked on his command to punch his fastballs up and in and especially against right-handed batters, as you can see in the below with 2017 on the left and 2018 on the right and with both charts being from the catcher’s perspective.

The end goal of these shifts in Kyle Freeland’s game was to cause as much soft contact as possible. As Freeland said, “Getting in on their hands is going to induce a lot of weak contact, especially if they aren’t able to get that barrel around and then once you do that, it opens up options to where you can throw your changeup down and away, and it comes out of your hand looking like a fastball, and then the next thing they know it’s off the end of their bat for a weak ground ball or a weak fly ball”. This concerted effort to induce soft contact not only proved to be fruitful for Freeland (who finished 18th in groundball percentage), but it seemed to be an effort pushed by the Rockies pitching staff that helped Jon Gray (10th) and German Marquez (12th) achieve similar results in 2018.

But what should we expect from Kyle Freeland in 2019? The vast majority of projections I’ve seen (PECOTA, FanGraphs, etc.) seem to suggest that Kyle Freeland’s 2019 season will be more similar to his 2017 season than his 2018 season. Considering that generating weak contact does not seem to be a pitcher skill that typically translates year over year, I can see why those projections are positioning Kyle Freeland as a major regression candidate. And given the historical performance of some of those projection systems, Freeland probably is very likely to show significant regression in 2019.

But I don’t really care, because Kyle Freeland fascinates me endlessly and I’ll root for him (and those unders) until we both fail.

The Evolution of Batting Metrics

Last week, I started off this year’s set of MLB write-ups with an introduction to cluster luck and the BaseRuns metric. This week, I thought it’d be best to expand on the offensive side of baseball and dive deeper into the evolution of batting metrics. The collective understanding of batting performance and efficiency has evolved over time and continues to see significant developments to this day. As a result, it can be difficult to determine which metrics are best for your own personal use when trying to model the offensive side of baseball. There is certainly no all-encompassing “right” answer when it comes to batting metrics, but there is certainly a great chance there is a “right” answer when it comes to finding metrics that tailor most to your own beliefs as to what should matter and how much it should matter.

Batting Average (AVG)

Batting Average is obviously the most widely-know batting metric there is, and is calculated by simply dividing total hits by total at bats. At best, batting average is a surface-level measurement of showing how often a player gets a hit (duh). The primary shortcomings of batting average are that it fails to quantify the plate appearances that don’t register as at bats (walks, sacrifice hits, etc.) and it fails to give any weight to the varying types of hits (a single and a home run are equally just one hit).

On-Base Percentage (OBP) and Slugging Percentage (SLG)

On-base percentage is exactly what is says it is and aims to address the first aforementioned shortcoming of batting average. OBP takes the batter’s instances of getting on-base (H + BB + HBP) and divides it by total plate appearances, and tells us the rate at which a batter gets on base. Slugging percentage aims to address the second shortcoming of batting average by weighing each type of hit by the number of bases a batter takes for each:
SLG = [ (1B) + (2B x 2) + (3B x 3) + (4 x HR) ] / AB
But although OBP and SLG each present a solution to one of the two major shortcomings of BA, they also each fail to address the remaining one. Which brings us to…

On-Base Plus Slugging (OPS)

OPS brings us the first official step into “sabermetrics” territory, first popularized in 1994 in The Hidden Game of Baseball by John Thorn and Pete Palmer. As the name suggests, OPS is calculated by adding OBP and SLG together, and aims to represent a player’s ability to get on base and hit for power. For your own reference, an OPS of ≥0.900 is typically considered to be a great mark to hit. One of the underlying problems with OPS lies more with the underlying problem with OPS, in that it linearly weighs the different types of hits according to the amount of bases they equate to.
A variant of OPS called OPS+ has since been developed that accounts for park factors and normalizes the stat across each of the two leagues (NL and AL). OPS+ is also scaled in a way where 100 is league average and each point of deviation above/below 100 equates to the percentage that the player is better/worse than league average. For example, a player with a 120 OPS+ is 20% better than league average whereas a player with a 85 OPS+ is 15% worse than league average.

Weighted On-Base Average (wOBA)

wOBA is a much more recently developed sabermetric, originally introduced in The Book in 2007.  The metric was actually developed and presented as an improvement from what OPS represented, as the authors felt OBP and SLG had significant overlap and that the on-base element of the statistic was being underrepresented. wOBA’s formula assigns “linear weights” that represent the average number of runs scored in a half-inning after such event occurs. These run value weights are then scaled to fit wOBA on the same scale as OBP (0.000 to 1.000). The formula for wOBA has evolved over time, with the first formula below being the original iteration and the following one being the Fangraphs version of the formula for the 2018 season.


Desipite the creators of wOBA hailing their newly-created sabermetric as superior to OPS in nearly every way, some research since has shown otherwise. In 2013, a professor from San Antonio College compared the predictive performance of OPS and wOBA and his results using the 2003-2012 seasons showed that OPS had a higher correlation to team run production rates than wOBA did. In 2018 Baseball Prospectus stepped in and conducted their own research, expanding the sample (1986-2016) and expanding the analysis to include a look at:

  • Descriptive performance: the correlation between the metric and same-year team runs/PA
  • Reliability performance: the correlation between the metric and itself in the following year
  • Predictive performance: the correlation between the metric and the following year’s runs/PA.

The findings essentially confirmed that OPS was superior to wOBA:

Runs Created (RC), Weighted Runs Created (wRC), and Weighted Runs Created Plus (wRC+)

The original Runs Created metric was created by Bill James and serves as an estimate of how many runs a player contributed to his team. Weighted Runs Created was an evolution of Bill James’ original work that incorporated the aforementioend wOBA into the formulation. The inherent problem with both itereations was that the stat was ultimately still just a counting stat, much like HRs or RBIs. Weighted Runs Created Plus (wRC+) did to Weighted Runs what OPS+ did to OPS, in that it took an otherwise context-less stat and turned it into a rate (while also controlling for park and league factors). Just like OPS+, a player with a wRC+ of 118 has contributed 18% more runs to his team than the league average player.

Wins Above Replacement Player (WAR or WARP)

The Runs Created trio above aren’t the only metric that tries to serve as an estimator of a player’s contributions to his teams offensive production. WAR is the number of wins a player has incrementally added to his team above the amount of expected wins if that player were to be replaced with a replacement level player. WAR as a whole incorporates batting, baserunning, and defense for position players, but the batting element of WAR can be singled out and is often represented as bWARP. The various baseball analytics sites (Baseball Prospectus, Baseball Reference, FanGraphs, etc.) have different formulations for WAR, so you may see some varying WAR figures.

Batting Average on Balls in Play (BABIP)

BABIP is another self-explanatory metric, as it essentially shows how many of a batter’s batted balls go for hits and outs. However, BABIP is much different than any of the stats discussed thus far. The primary purpose of BABIP is to serve as a potential warning that a batter is possibly performing above (high BABIP) or below expectation (low BABIP). In essence, a batter with a BABIP with significant deviation from the “normal” .300 typically signals that the batter is due for regression towards the mean. That regression expectation can be applied within a season or from year to year. To show BABIP in action, here is a chart showing the BABIP leaders from 2017 (minimum 250 plate appearances) and their performance in 2018. Significant increases in performances are highlighted in green, significant decreases are highlighted in red, and relatively similar performances are left uncolored:


BaseRuns (BsR)

BABIP begins to lean into the world of expectation and the “should’ve could’ve would’ve” world of sabermetrics, and that is arguably the largest development to date in the space. We covered BaseRuns quite in depth last week but as a refresher, BaseRuns is a metric that was originally developed by David Smyth and aims to estimate how many runs a team should have scored over the course of a season. Since we played with BaseRuns a lot in the previous write-up, I won’t expand any further in this write-up.

Deserved Runs Created Plus (DRC+)

Deserved Runs Created Plus is the newest development in batting sabermetrics, having just been introduced by the Baseball Prospectus team this past December. I could try my best to explain the premise of DRC+, but Baseball Prospectus did exactly that in their aptly titled “Why DRC+?” article that accompanied the introduction of the metric:

“Why another batting metric? Because existing batting metrics (including ours) have two serious problems: (1) they purport to offer summaries of player contributions, when in fact they merely average play outcomes in which the players participated; and (2) they treat all outcomes, whether it be a walk or a single, as equally likely to be driven by the player’s skill, even though no one believes that is actually true. DRC+ addresses the first problem by rejecting the assumption that play outcomes automatically equal player contributions, and forces players to demonstrate a consistent ability to generate those outcomes over time to get full credit for them. DRC+ addresses the second problem by recognizing that certain outcomes (walks, strikeouts) are more attributable to player skill than others (singles, triples).”

Like the rest of the “plus” metrics, DRC+ is scaled in a way where 100 is league average and deviations above/below signal how much better/worse a player is in terms of percentage. DRC+ was met with some criticisms following its unveiling, and the Baseball Prospectus team has since updated the metric in response. The team also did extensive research to compare the updated DRC+’s descriptive, reliability, and predictive performance compared to OPS+, wRC+, and the original DRC+:


Bringing It All Together

For your reference, above is a chart showing how many runs each team scored in 2018 as well as how they performend in each of the batting metrics we covered with the exception of WAR (too many variations to choose from) and BABIP (not really purposeful for this chart). It’s easy to see how our starting point of batting average relatively fails to be an accurate measure of batting production and/or efficiency, as team performance in that category has the lowest correlation to runs scored than any other metric represented. More importantly, this chart illustrates that most batting metrics will give you a similar general idea of a team’s batting ability, showing that each metric typically has some value when it comes to evaluating batting performance.

And that is essentially the abridged version of the evolution of batting metrics. Hopefully this gives you a better understanding of what it is that you’re exactly looking at the next time you pull up a stats page on Baseball Prospectus or FanGraphs. As always, if you have any questions about the topics covered in this write-up you are more than encouraged to reach out to me via Twitter.

Until next time.

Cluster Luck and BaseRuns

It is officially time for us to set our eyes on Major League Baseball. From now until the end of the season, I will be providing a weekly baseball write-up (hopefully) every Wednesday. Between now and Opening Day, I thought it’d be best to cover some introductory (yet still very comprehensive and higher level) topics, metrics, and ideas that populate the analytical side of professional baseball. For today’s write-up in particular, I will be answering the question “Is it better to be lucky than good?” by taking a look at cluster luck and BaseRuns.

Cluster luck is a term coined by Trading Bases author Joe Peta that serves to be the underlying explanation as to why the amount of games a team actually won would differ from the amount of games they should have won. Cluster luck itself is not an actual metric with a formula, but there are plenty of ways to calculate expected wins and see cluster luck in action. The Pythagorean win theorem, a formula created by Bill James (the founding father of sabermetrics), estimates the percentage of wins a team should have had given the amount of runs they have scored and the runs they have allowed. The original formula was:

(Runs Scored)^2  /  [(Runs Scored)^2 + (Runs Allowed)^2]

Since the original formula was published, sabermetricians have more accurately assigned 1.83 as the exponent. Putting the formula to action, the World Series-winning Red Sox scored 876 runs and allowed 647 runs. Plugging those into the modern iteration of the Pythagorean expectation formula with 1.83 as the exponent, the Red Sox should have had a 0.635 win percentage, good for 102.9 wins. The Red Sox finished the regular season with 109 actual wins, a difference of 6.1 wins. Generally speaking, a four win difference between actual wins and Pythagorean expected wins is considered to be significant and generally non-repeatable. In other words, the 2018 Boston Red Sox very likely benefited from cluster luck.

But what exactly is “cluster luck”? Cluster luck is the idea that the particular sequencing of plate appearance outcomes lead to very different run-scoring and run-allowing results. Given that each plate appearance by any given player can be numerically boiled down to a set of expected probabilities assigned to each possible event, pure chance has the ability to cluster positive or negative events sequentially thus leading to very different outcomes. Take an inning in which a team has two walks, one single, one triple, two strikeouts, and one popout. Depending on the sequencing of those events, you can have two very different outcomes:

  • Sequence A: Single, BB, Strikeout, BB, Strikeout, Triple, Popout
  • Outcome A: Three runs scored / allowed
  • Sequence B: Triple, Single, BB, Strikeout, Strikeout, BB, Popout
  • Outcome B: One run scored / allowed

Sequence A would be considered to have an outcome that the offense benefited from cluster luck, whereas Sequence B would be considered to have an outcome that the defense benefited from cluster luck. Throughout the course of 162 games, teams can become significant victims or beneficiaries of cluster luck. Here is a table showing the lucky teams and the unlucky teams of 2018:

s always (and especially with baseball), we have an ability to go deeper with our evaluation. In particular there is the BaseRuns (BsR) metric which was originally developed by David Smyth and aims to estimate how many runs a team should have scored or allowed over the course of a season. Much like the Pythagorean expectation formula (and the vast majority of sabermetrics), the BaseRuns formula has evolved over time and there are some slight variances depending where you look, but the current iteration I utilize is the FanGraphs version that currently looks like this:

The above table shows BsR scored with the “lucky” teams on the left having more actual runs scored than BaseRuns scored, and the “unlucky” teams having the opposite. We can use the above table to potentially begin answering the “Is it better to be lucky than good?” question. Of the top nine teams in offensive BaseRun differential (actual runs minus BaseRuns), six played in the division tiebreaker games or made the playoffs outright. On the other hand, ten of the bottom eleven teams in differential missed the playoffs. BaseRuns alone are also just a solid measurement of a team’s strength in run production, as last year all ten teams who played in a division tiebreaker or made the playoffs outright finished in the top thirteen in BaseRuns scored. Next, let’s look at BsR allowed:

Obviously the inverse would be true with BsR allowed with the “lucky” teams on the left having less actual runs allowed than BaseRuns allowed, and the “unlucky” teams having the opposite. Five of the top eight teams that have a benefitial differential made the division tiebreakers or playoffs outright, whereas eight of the ten teams with the most unfortunate differentials missed the playoffs. Nine of the ten teams who made it that far also ranked in the top twelve in least BaseRuns allowed. The Rockies are the only such team to miss that cut, but given that they play in run-friendly Coors Field it’s easy to understand why they might not rank favorably in BaseRuns allowed. Nevertheless, now that we have an expected runs scored metric and an expected runs allowed metric, we can use the Pythagorean win formula to see how many wins each team should have won in 2018. Below shows exactly that, with the left table being sorted by differential and the right table being sorted by Pythagorean expected wins using BaseRuns. I’ve also highlighted the ten teams that played past Game 162.

As you can see in the right table, Pythagorean expected wins seems to be a good measure of team strength with nine of the ten “postseason” teams ranking in the top eleven. The Rays might be the team that suffered the worst cluster luck fate given the circumstance of their misfortune. They finished fifth in expected wins but finished eleventh in actual wins. Furthermore they finished seven games behind the Athletics for the final Wild Card spot. If you were to bring them from -5.67 to 0.00 in Pythagorean vs. actual win differential, it would have put the Rays at 95.67 wins. If you do the same for the Athletics with their differential (moving them from 2.49 to 0.00), the Rays would have had a better record and beaten them out for that Wild Card spot. On the other side of the coin, the 91 win Rockies certainly benefited the most from cluster luck as they should have had 84.39 expected wins using BaseRuns, which would have been bested by either the Nationals or the Cardinals when bringing their expected vs. actual win differential to zero. So to answer the question, yes, it is in fact better to be lucky than good (sometimes).

That’s going to wrap it up for my inaugural MLB write-up. I hope this was an insightful and educational start to what is hopefully a very insightful and long-lasting series of write-ups for the MLB. If you have any questions regarding the topics I covered today or have any topics in mind that you would like me to cover in future write-ups, don’t be afraid to give me a shout!

Until next time.

NFL Model Recap

With the Super Bowl behind us, the 2018 NFL season comes to an end. This season definitely was quite the ride, and the debut of my NFL model concluded with a 75-47-6 record (good for a 60.94% hit rate) and a +89.30 unit yield and 18.10% ROI. Before we say goodbye to the NFL season, I thought it’d be interesting to do a model recap and take a deeper look at how the model performed across various splits and to highlight some things I’ve eyed as potential areas of improvement for the 2019 season. Let’s get started.

Did higher unit plays generate higher profit?

One of the biggest ways to measure the strength of a model is to see how it performs as its disagreement levels increase. I first took a look at model plays of 3.5 units or more, as there doesn’t exist a spread that doesn’t move off of or through a key number or zero when adding 3.5 points. I then went up by two units from there. Here’s a look at those splits:

  • All plays: 75-47-6 (60.94%) for +89.30 units (18.10 % ROI)
  • 3.5+ unit plays: 34-21-0 (61.82%) for +52.94 units (19.04% ROI)
  • 5.5+ unit plays: 10-6-0 (62.5%) for +22.19 units (21.36% ROI)
  • 7.5+ unit plays: 3-2-0 (60.0%) for +9.53 units (23.59% ROI)

As you can see, the ROI steadily increases as the unit split increases. This is very reassuring for the model as it demonstrates the model’s ability to scale unit disagreement with profit expectation. On top of that, the fifteen largest plays of the season had a 10-5-0 record with a 27.82 unit yield, which was good for a 28.32% ROI. This obviously includes a win on the model’s largest play of the season when the Panthers miraculously covered a 8.8 unit play on a +6 spread in their Week 15 matchup against the Saints. That also marked the model’s peak profit point of the season at +102.27 units. Furthermore, the top 15 model plays were all on underdogs, and 37 of the top 38 model plays were on underdogs. 26 of the top 29 model plays were also on home teams. Here’s a look at the overall home / road and favorite / underdog splits:

How did the model do by team?

Speaking of the Panthers, they happened to be the most profitable team to bet on for the model this season and the third-most profitable team to bet against. Across those two sets of plays for the Panthers, the model went 7-0 for +28.52 and an insane 98.32% ROI. Below is a look at the earnings betting on each team, betting against each team, and total earnings on  model plays relating to each team.

Looking at the last chart, there are only 11 teams the model had overall losses on (only 9 were of >1 unit) while it managed to profit on the remaining 21. Outside of that, there are definitely some other key takeaways you can pull from this. Looking at the “Earnings Betting Against Each Team” chart, you’ll see that 11 of the 17 teams that the model profited on were teams with winning records. Out of the other six teams, three made the playoffs last year (CAR, ATL, JAX) and one is a historical public favorite (GB). This demonstrates the typical overvaluing of good teams by the public and the consequent overpricing of those teams in the market to capitalize on that perception. When looking at the “Earnings Betting On Each Team” chart, you’ll see that 8 of the 11 most profitable teams had losing records. The counterpoint would be that 7 of the 8 teams that are net negative in that chart had losing records as well, but the losses of those seven teams are offset by the earnings generated by just the top five profit-generating teams with losing records.

Was your low juice tolerance +EV?

Many, especially those who read these newsletters, have noticed that I try my best to keep the juice low on model plays (-105 or lower). I succeded in that regard as 80.5% of all model plays had a juice of -105 or lower. A common misconception was that I was often selling a half point, which was definitely not the case. This is because the book I used for this experiment had a reduced juice of -105 on equal lines as opposed to the “standard” -110. However, there are 33 model plays out of the total 128 plays (25.6%) that had a juice of -102 or lower. Of those 33, two plays lost by a half point and one play pushed. They were:

  • Loss on TB -3.5 -102 (2.254 for 2.2u) Week 7 vs. CLE
  • Loss on NYJ +6.5 -101 (5.781 for 5.7u) Week 15 vs. HOU
  • Push on SEA +2 +106 (2.1 for 2.218u) Wild Card round @ DAL

The first play wasn’t really selling a half point since you pay way more juice to move to a key number. TB -3 would have likely been around -120, which is just way too much juice. The same applies to the second play, as a NYJ +7 line would have similarly been around -120. On top of that, I removed spreads of +7 and higher from model consideration by that point so NYJ +7 wouldn’t have been a play for that reason as well. However, the last play I did indeed sell a half point and it was very lucky to push but I still stand firmly behind the proposition of selling points, especially around non-key numbers. Speaking of key numbers…

Are “key numbers” really that important?

These splits were actually the most eye-opening to me, as I really didn’t take a good look at it until the season was over:

  • Plays on +3 and +3.5: 20-7-1 (70.69%) for +37.05 units (38.07% ROI)
  • Plays on +2.5: 2-5-0 (40%) for -6.79 units (-25.90% ROI)
  • Plays on -2.5 and -3: 5-3-1 (61.11%) for +6.07 units (25.47% ROI)
  • Plays on -3.5: 1-3-0 (25.00%) for -4.74 units (-52.56% ROI)

I didn’t include plays on or around -7 and +7 since there weren’t many plays at all on those numbers, especially with the model exclusion that was established Week 10 onward.

What are you looking to change for 2019?

As great as the model’s performance was both on a macro and micro level, there are certainly many improvements I am looking to make in preparation for next season. The most obvious areas of improvement would be taking a look at the two things that caused enough trouble to warrant on-the-fly model exclusions: Thursday Night Football games and spreads of +7 and higher.

For Thursday Night Football, I don’t believe there is a mathematically derivable solution or adjustment that I can make that would provide a consistent positive performance. The game of football is just played so entirely different when you have just three days to rest and prepare, and for that reason I fully anticipate removing Thursday Night Football games entirely from model consideration from the start.

As for spreads of +7 and higher, I will have to take a more thorough and refined look at the win probability data at the more extreme handicap levels and see what adjustments I can make to more accurately capture edges in those ranges. I think the relatively poor performance on such spreads may have been tied to the league-wide jump in scoring, but that will obviously take some time to investigate and even if my assumption is true, using a single season of data would be a dangerous proposition.

Another thing I’ve mentioned previously as something I’d like to take a deeper look at in the offseason is the entire “not every point is created equal” reality that comes with football spreads. This applies in several ways, whether it be points moving on/off key numbers being worth more or points being worth more in low-scoring environments and less in high-scoring environments.

I’ve also considered backtesting how the model would have performed if all model plays were done on the moneyline as opposed to against the spread. This could potentially be beneficial as it avoids the “not every point is created equal” dilemma while also aligning our desired outcome (a win) with the team’s desired outcome (a win).

On the complete opposite end of the spectrum, I want to see if there is a way to backtest how the model would have performed if model plays consisted of playing alternate spreads that matched the model’s spread. This would be interesting as the disagreement level between the model expectation and the Vegas implied probability would be captured in the odds of the alternative spread, and the staking method would just have to be a flat unit on each wager.

All in all it was a great season, whether it was the model’s performance or the amazing response I’ve gotten from all of you. I’d like to thank you all again for your support and interest in my work. I am more than excited to see how my work with other sports compares, especially with the MLB. Beginning next week I’ll be pivoting my weekly write-ups to the MLB, covering several higher level analytical topics as we inch closer and closer to Opening Day.

And if you need just a little more NFL content to hold you over, I did join Whale Capper and Andy from the Deep Dive podcast to discuss some of the lessons we’ve learned from our 2018 NFL seasons. I think we had some pretty insightful discussion on quite a few topics, so make sure to check it out! And as always, don’t hesitate to reach out to me via Twitter.

We’ll talk again soon.

NFL Playoffs: Super Bowl LIII

It’s been quite a while since we last spoke. Both of the conference championships had some unfortunate events not fall our way, namely a missed pass interference / helmet-to-helmet call in the NFC Championship game that compromised our NO -3 play, and then an unfortunate neutral zone infraction by Dee Ford that negated a game-sealing interception that would have won us our KC -3.5 model play. Despite those worst-case scenarios, playoff model plays sit at 3-3-1 for basically a wash at +0.03 units. More interestingly, those two losses on home favorites moves that split to a negative ROI on the season – the only split to do so across home / road and favorite / underdog splits:

But enough about that, there is now just one game remaining in the 2018 NFL season. It’s been a long journey to get to this point for both the Patriots and Rams, and it would only be right for us to take a deeper look at both of these teams and this monstrous matchup ahead of the biggest sports game of the year. There’s quite a lot to unpack, so let’s get right into it.

#2 Patriots (13-5) vs. #2 Rams (15-3)

I’m going to skip right past all of the narrative stuff that I’m sure you’ve been bombarded with the last two weeks (SB 36 rematch, young vs. old coach/QB, etc.) and dive head-first into the numbers. Before the season began, the Rams were projected to be the second-strongest team at 10.68 expected wins and the Patriots were projected to be the seventh-strongest team at 9.66 expected wins. As the season progressed, the Rams continuously outperformed their expectation while the Patriots somewhat tread water in the eyes of the model. The culmination of this was a gap between the two that peaked at a 2.91 expected win difference. The Rams then spent the remainder of the season tumbling back down to Earth, landing at 10.53 expected wins and within just 0.15 expected wins of their preseason projection. Meanwhile, the Patriots continued to tread water and would go on to finish the season with 9.86 expected wins, which was just 0.20 expected wins off from their preseason projection.

Last write-up I spent a lot of time talking about Rams defensive coordinator and his history of devising defensive gameplans to stifle elite quarterbacks and passing attacks. This week, he of course has the task of slowing down Tom Brady. Not enough can be said about Brady, especially in regards to his efficiency against blitzes even in his “old age”. Since 2016, the Patriots are 29-2 when brady has been pressured less than 27% against four or less pass rushers. The quick answer a lot of defensive coordinators come up with when faced with a statistic like that is “blitz more then”. But during that same stretch, the Patriots are 15-1 when Brady is blitzed 30% or more on dropbacks (via NextGenStats). So what is the appropriate gameplan?

I think Wade Phillips knows that answer. The last time Phillips faced off against Brady and Belicheck in the playoffs (2016 AFC Championship), he crafted a defensive gameplan that limited Brady to a 48% completion rate while forcing two interceptions and generating four sacks. He did so by utilizing the gameplan I highlighted two weeks ago in my newsletter: utilizing line-of-scrimmage disruption on the pass catchers to delay route development in order to generate an effective four-man pass rush. We’ve seen throughout these playoffs that Brady has been able to completely neuter solid pass rushes by getting the ball out quickly. Forcing Brady to hold the ball for an extra second makes a world of difference, as it gives Aaron Donald (who has the fastest time-to-sack average in the league) and Ndamukong Suh more time to penetrate the pocket. Only sending four to generate pressure gives the Rams more resources to dedicate in coverage, making Brady’s job even harder.

Despite all of that, Brady will not be the only quarterback who could have a hard time come Sunday. In the Patriots two playoff games thus far, they’ve generated two of their three highest pressure rates of the season (45.3% against the Chargers and 44.4% against the Chiefs). Unlike Brady, Goff isn’t quick to get rid of the ball. In fact Goff holds the ball on average for 2.96 seconds, the fourth-longest of all starting quarterbacks this year. That characteristic of Goff’s is blood in the water for the Patriots pass rush, as the Patriots have secured a perfect 10-0 record this year against quarterbacks holding the ball 2.8 seconds or longer (via ESPN Stats & Info). The one caveat to all of this is that this tendency of Goff’s is a product of the Rams’ reliance on play-action passes, which only exists due to the quality and outstanding performance of their offensive line. It obviously remains to be seen who will win that battle in particular.

If the Rams’ offensive line happens to hold up against the Patriots pass rush, then the Patriots coverage scheme become arguably the single-most important element to this game. As mentioned in a previous newsletter, the Patriots utilized man coverage more than any other team in the NFL this season. It would not surprise me one bit if Patriots Defensive Coordinator Brian Flores and of course the defensive-minded Bill Belicheck opt to rely less on man coverage and utilize more zone coverage. The reason being is that Jared Goff has the highest yard per attempt average in the league against man coverage this year, and has thrown 16 TDs and just 2 INTs against man. The magnitude of that efficiency is more apparent when compared to his 9 TD / 9 INT split against zone coverage.

The model has an easier time projecting the Super Bowl than it does any other game because the game is played on a neutral field, requiring no travel or homefield adjustments. This particular Super Bowl also has an extremely clean injury report, meaning the data is even more pure than usual. As for the actual numbers, the model actually projects the Rams as the favorite with a 51.64% chance of winning the game which converts to a LAR -1 model spread. The spread has been stuck in the mud at NE +2.5 for the past ten or so days and I don’t anticipate a move to LAR +3 that comes with juice of -110 or lower. LAR +2.5 would be an official model play for 3.5 units but the juice unfortunately moved from -102 to -109 between the time I started this write-up and the time I submitted it to be published. I do think I can get better than the current -109 juice and will wait to do so, and as always you can expect a tweet from me with the official model play when it is made.

And with that, the NFL 2018 season as well as the accompanying model plays come to a close. I hope my write-ups this season provided some insight into modeling the NFL and how to assess matchups. I want to thank all of you who have reached out via Twitter with words of encouragement, it certainly means a lot. Next week, I’ll return with a full recap and in-depth look at the model and its performance. This will include week-by-week splits, team splits, and areas of potential improvement for the 2019 season. If you have any questions or particular things you would like me to cover, make sure to give me a shout on Twitter.

And with the NFL season coming to a close, I now have my eyes set on the 2019 MLB season and the model I’ve been building in preparation for it. Following next week’s NFL model recap write-up, I plan to do a weekly MLB series that will cover higher-level MLB topics and model principals, so stay tuned for that. The plan as of right now is to provide MLB model plays for free through April, and to utilize a publicly available Google Sheet that allows you guys to see what the model line is and at what prices each team begins to show enough value to be a model play. I have some other tools in the works as well, so make sure to keep your eyes open for all of that.

Thanks again for reading.

NFL Playoffs: Conference Championships

The Divisional round saw a split between the two model plays of KC -4 and LAC +4.5, bringing the model to 75-44-6 for a +93.67 unit yield and a 19.98% ROI while playoff model plays move to 3-1-1 for +4.41 units. In last week’s pair of newsletters, I wrote about the numerical value of byes and the numerical effect of weather to help give us a better grip on what was driving last week’s spread numbers. This week, I’d like to discuss a game between the Green Bay Packers and Denver Broncos from 2015 and the significance it holds on the two Conference Championship games being played this week. No I am not kidding, but I am aware that sounds strange so I’ll jump right into it and explain its significance.

November 1, 2015 is a night I’ll remember for as long as I’m a football fan. It was Week 8 of the NFL season, and the 6-0 Packers who had just come off the choke job of the century against the Seahawks in the NFC Championship the previous season were travelling to Sports Authority Field to take on the 6-0 Denver Broncos on Sunday Night Football. As a Packer fan, the 6-0 start was reassuring with a win at Soldier Field, a revenge win against the Seahawks, and a win on the road against Colin Kaepernick and the 49ers who had bounced the Packers in the 2012 and 2013 playoffs. Aaron Rodgers was essentially untouchable, boasting a 76-37 (67.3%) record as a starter heading into that game. I was feeling great about the Packers.

And then we got absolutely romped. The Broncos went up 17-0 by the middle of the second quarter and I knew it was over. I didn’t care that we had Aaron Rodgers, who seemingly could make a game out of any of them no matter what the deficit. This game just felt different; it felt like Broncos Defensive Coordinator Wade Phillips had figured out the formula on how to beat Aaron Rodgers and the Packers offense. That offense had beaten teams for years as many defenses relied on soft man and zone coverage with plenty of help over the top to defend the deep ball, or with over-committed blitzes to try and generate meaningful pressure. Wade Phillips on the other hand had the talent and gameplan that finally did what so many had failed to do until then.

The Broncos’ gameplan involved generating an effective pass rush with just four men while playing press man coverage. With Jordy Nelson going down with a season-ending injury in the preseason, the Packers’ really lacked a pass catcher with the skill set to beat disruption at the line of scrimmage. The delay in route development caused by that disruption was long enough to give the Broncos’ pass rushers just enough time to generate meaningful pressure, wreaking absolute havoc on Rodgers’ ability to find an open man. Rodgers would finish the game 14 of 22 for 77 yards and a 65.8 rating as the Packers lost 10-29. The Packers would finish that season 9-7 and with an exit in the Divisional round of the playoffs. Rodgers since that game has gone 24-23-1 as a starter (51.0%). Wade Phillips and the Broncos defense would go on to become the centerpiece of a masterful season that culminated in a Super Bowl win. So what does this have to do with this week’s games? Let’s get right to that.

#2 Rams (14-3) @ #1 Saints (14-3)

Wade Phillips, now the defensive coordinator of the Rams, will have a similarly tall task to the one he faced for the November 1st, 2015 game against Aaron Rodgers. This time the stakes are quite a bit higher as he will be going up against the Saints in the NFC Conference Championship, a team that ranks third in pass offense and eigth in run offense. In their first meeting, it’s safe to say that Wade Phillips and the Rams’ defense utterly failed, allowing 45 points, 487 total yards, and a 346 yard – 4 TD – 0 INT line from Drew Brees while totalling zero sacks. Although they picked up a decisive win last week, the Rams only sack came on a strange play where forward progress was deemed stopped as Prescott was being held up and pushed forward by his own lineman. To add to the disappointment, the Cowboys offensive line ranked 27th in adjusted sack rate this season. To even further add to that disappointment, the Rams played that game at home in LA which means crowd noise was at its peak as the Cowboy’s o-line was waiting for the ball to snap. The tables could not be turned any more for this week as the Saints rank third in adjusted snap rate and this game will of course be played in the raucous Superdome.

To draw another parallel to that aforementioned Packers-Broncos game, C.J. Anderson lit the Packers defense on fire in that 2015 game with 101 yards on 14 carries and a touchdown (while having two goal-line TDs sniped by Ronnie Hillman). Last week, CJA continued his recent renaissance with the Rams with a 23 att / 123 yds / 2 TD line while teammate Todd Gurly added on a 16 att / 115 yds / 1 TD performance. However, the Saints’ third-ranked rush defense comes more than equipped to slow down the two backs as it held Eagles’ running backs to 37 yards on 13 carries last week. Even if we ignore the Saints’ ability to defend the run, I think the Rams’ backfield will be plenty busy in pass protection anyways as the Saints rank fifth in pressures generated in home games (via Sports Info Solutions).

In their first matchup, Sean McVay found only 13 opportunities to hand off the ball to his backfield and I expect much of that to hold this time around. However, passing more for the Rams could actually play to their favor, as the Saints are allowing a 47% success rate on pass attempts against 11 personnel (which the Rams run a ton of), a number that ranks bottom 27th in the league (via Sharp Football Stats). On top of their defeciencies defending the pass against 11 personnel, the Saints also struggle incredibly against deep passes. In fact, the Saints are the worst team in the league at defending deep passes (>15 yards) according to Football Outsiders’ DVOA metric.

As for the model, this game is made to be NO -5.1 which means a model play on the Saints would begin at -3 whereas the Rams would become a model play at +7.5 (which we have no chance of seeing). I do see NO -3 currently but it is priced at -114, which would be a good amount higher than I could tolerate. If and when the juice comes within an acceptable range, I’ll be sure to lock in the Saints as a model play on Twitter.

Much like the 2015 Broncos, the Patriots utilize a lot of man coverage. In fact, they led the league this year in man coverage usage with 56.8% of their snaps being in man (via Sports Info Solutions). But unlike the Broncos, the Patriots may need to shy away from man coverage in their defensive game-planning if they want to succeed. Patrick Mahomes has played five games against teams in the top twelve of man coverage usage and in those games he has averaged 328 passing yards, 3.8 TDs, and a 69.2% completion rate, which are all above his already absurd season averages. Sammy Watkins, who just returned last week from a lengthy injury, has averaged 4 rec / 74.8 yds / 0.5 TDs in those particular games. Watkins first season with the Cheifs has proven to be a success when he has been able to play, as he has averaged 5 rec / 64.1 yds while generating a 120.0 passer rating when targeted and a 72.7% catch rate on a 80% catchable target rate, all of which are either career highs or second to his career high (via Graham Barfield).

Although the Patriots have a run defense that otherwise grades as average overall, they do have one glaring weakness that went unchallenged in last week’s game. The Patriots are horrid at defending runs out of 11 personnel, allowing a league-worst 61% success rate and 6.8 yards per carry. The Chargers, who have never really done much to show that they care about or gameplan around analytics, gave Melvin Gordon just four carries last week from 11 personnel. On the other hand, Chiefs running back Damien Williams runs from 11 personnel on 64% of his attempts while averaging a 67% success rate and 5.2 yards per carry.

It’d be cheeky to talk about the Patriots’ defensive shortcomings given the defensive unit they’re going up against this week. To say that the Chiefs have a poor defensive reputation would be an understatatement. On the surface, they allow the ninth-most points per game and the second-most yards per game. That is certainly not good, but I do think there are certain elements of their defense that could pose some trouble for the Patriots. The most notable is that their front seven has generated the most quarterback pressures in home games this year. Granted the strength of the opponents they’ve played at home isn’t the greatest, but when looking at pass defense DVOA (which is opponent-adjusted) the Chiefs rank twelvth in the league. They’ve also played three playoff teams in their last four home games, and in those games they’ve allowed an average of just 221 passing yards while surrendering a 59.2% completion rate. Those games included games against the second-ranked Chargers’ pass offense and tenth-ranked Colts’ pass offense. They also generated ten sacks against those offensive lines, which ranked 2nd, 8th, and 13th in adjusted sack rate. To give you an idea of how impressive that 3.33 sacks per game average is, the league leader in sacks per game was the Steelers with 3.2 sacks per game.

Speaking of home and away splits, those get even uglier the deeper you dive into the game. The Chiefs allowed the second-most points per game on the road (34.6) while allowing the third-least at home (17.4). Of those three aforementioned recent home games against playoff teams (BAL, LAC, IND), they allowed an average of 22 points. The Patriots splits are equally massive: their net yards per play was +0.9 (2nd) at home and -0.6 (26th) on the road, and their average margin was +15.9 (1st) at home and -2.4 (17th) on the road. And I’m sure many of you have heard many times this week that the Brady/Belichick era Patriots are 3-4 on the road in the playoffs (which accounts for half of their total playoff losses), but that can be sliced and framed so many different ways to play for or against them. Three of the losses are in Denver, a notoriously tough place to play with the included altitude element. Three of the losses are to some guy named Peyton Manning. The three wins are against the Ben Roethlisberger Steelers and the Philip Rivers Chargers, both of which have been manhandled particularly egregiously by Brady and Belichick.

But I don’t want to make it sound entirely doom and gloom for the Patriots. They obviously stand a reasonable chance to win this game and possess advantageous matchups that can help push the game more in their favor. The Patriots rely heavily on 21 personnel (two back sets), ranking second in the league in play frequency on the year at 29% and only trailing the 49ers who use it on 41% of plays and are the only team who use it as their most common personnel package. However, the Patriots have used 21 personnel even more since their bye and since their committee of running backs have returned to full health. Since Week 12, the Patriots have used 21 personnel on 35% of all of their plays. This could spell disaster for the Chiefs’ defense as they are the worst in the league at defending 21 personnel, allowing a 65% success rate overall and a 98.2 passer rating on pass plays and 6.2 yards per carry on run plays.

As for the model, the game is made to be KC -5.7 which means there would be a model play on the Chiefs beginning at -3 and a play on the Patriots at +8.KC is currently -3 -116 which is too much juice, but just like the Saints play I will make sure to share if and when that comes within an acceptable range.

I hope you enjoyed this week’s write-up! That will wrap it up for this week, but I’ll be back in your inboxes in two weeks’ time to discuss Super Bowl LIII. I hope for my sake that I’m writing about a game that involves the Saints.

  • 1/4: Saturday Wild Card
  • 1/5: Sunday Wild Card
  • 1/11: Saturday Divisional
  • 1/12: Sunday Divisional
  • 1/19: Conference Championships
  • 2/2: Super Bowl

Thanks again for reading!

NFL Playoffs: Divisonal Round Sunday

In yesterday’s write-up, I covered the playoff bye dynamic and what historical effects it has had on team performance. Today, I’d like to take a look at another game element that has a measurable effect on games and is also prevalent in this week’s slate of games. That effect is weather. Each of the three games being played outdoors this weekend have a weather element to them, and everyone from bettors to players to analysts to talking heads have chipped in with their thoughts on how each condition will effect each team and game. But as always, I like to slice right through the noise and use a data-driven approach.

Whether Weather Has An Effect

Let’s first take a look at temperature. In a study conducted by advancedfootballstats.com, teams were separated out by four climate types: warm, moderate, cold, and domed. Those teams’ performances as a road team were then analyzed in temperature increments to see the correlation between temperature and win percentage, and below are the results of that analysis. As you can see, the most discernible effect temperature has is on domed teams playing in cold weather. The rest of the results are what I would consider to be inconclusive, given that warm teams had the highest win percentage at extreme lows (11-20 °F) and cold teams had the best win percentage at extreme highs (81-90 °F). That would obviously go against the what you typically hear from just about everyone.

Next up is precipitation, which is obviously split into games with rain and games with snow. The former typically has a negligible effect on games, which may go against common narratives you hear. A fantastic analysis done by Chris Allen from 4for4 showed that when compared to a “clear” game (60-75 °F, <9 mph wind), rain games have just a 2.4% decrease in total pass attempts, a 2.0% decrease in air yards, and a 0.5% decrease in deep pass attempts. Even target distributions by position don’t see a noteworthy enough change, which you can see below. Before tackling snow’s effect, I thought I’d quickly mention that games with significant wind see a negligible 2.4% decrease in total pass attempts and a 0.8% decrease in air yards, but a most-certainly noteworthy 6.2% decrease in deep pass attempts and a 4-8 point decrease in actual game totals when reaching 15 mph.

Snow is a weird weather element when looking at the data. The common narrative with snow is that passing volume decreases and rushing volume increases, which is generally true with a 6% swing in pass / run ratio. But what about the effect on efficiency of those types of plays? In the aforementioned 4for4 analysis, there was an 8.3% drop-off in total pass attempts, yet just a 1.4% decrease is deep pass percentage and a 5.3% increase in air yards. But not all snow is the same. In a study done by Pro Football Focus, it was found that “light” snow has a very negligible effect on passing efficiency (1.8% increase in completion percentage) and rushing efficiency (0.27 more yards per carry). It isn’t until snow is “heavy” that the there begins to be a more profound effect, with a 8.4% dip in completion percentage being the most notable change. I believe that public bettors tend to overreact at just the very sight of snow, and you see this with the light snow expected to be on the field for the Colts-Chiefs game despite the fact that it isn’t expected to snow at all during the actual game.

#5 Chargers (13-4) @ #2 Patriots (11-5)

The Patriots are a team the model can channel their Dennis Green with and say, “They are who we thought they were”. Coming into the season, the Patriots were expected to regress just a tad. They finished the 2017 season as the #3 team according to the model and they were projected to finish the 2018 season sixth-best with 9.66 expected wins. They ended up finishing the 2018 regular season with the fifth-best expected wins total, with 9.86. The Patriots never dipped lower than 9.21 expected wins and peaked at 10.03, showing that their performance has been consistent. Known to draw a lot of public money, their prices are often inflated and overvalued and the model finding zero times throughout the season to back them supports that. As for betting against the Patriots, the model found plenty of opportunities and went 5-2 for +13.36 units and a 45.01% ROI if we ignore the Week 5 model play backing the Colts against the Patriots on Thursday Night Football which remains the only model play I’ve personally advised against backing.

So what reasons would there be to back the Patriots this week against the Chargers? The model doesn’t really find any across the board. The Patriots rank behind the Chargers in every model category except special teams, which accounts for roughly just 6% of total team strength. Granted, the Patriots trail the Chargers in pass offense, rush offense, and pass defense by just a combined nine spots across all three categories, but it’s the Patriots 18th ranked rush defense going up against the Chargers’ sixth-best rush offense that will pose a problem in this game.

If there’s one thing the Patriots do have going for them, it’s that their defensive personnel and typical gameplan fits the mold of what brings down Philip Rivers’ effectiveness. The Patriots certainly have the ability to play man coverage, which yields 1.55 less yards per attempt from Rivers compared to zone coverage. The Patriots also pressure opposing quarterbacks at the fifth-highest rate in the league, which limits Rivers’ ability to wait in the pocket and fire deep to Mike Williams who may be drawing favorable coverage in this matchup if the Patriots’ top corner Stephon Gilmore is tasked with covering Keenan Allen.

Hopefully most of you have been following along on Twitter and received the play as soon as it was made, as I am only seeing LAC +4 available now, which would reduce the points of disagreement to 1.7 and no longer be enough for a model play.

#6 Eagles (10-7) @ #1 Saints (13-3)

The Eagles and Saints are two teams I’ve covered in detail in past newsletters, so I’ll just quickly recap their model numbers. The Eagles stand 15th overall in the model with 8.26 expected wins whereas the Saints sit fourth overall with 10.29 expected wins. An argument can be made that the Saints are actually undervalued in the model, as their adjusted Pythagorean win total for this year is 11.45 and of the remaining playoff teams, only the Colts possess a larger gap between the two numbers. Despite this game featuring the largest spread of the playoffs thus far, there is quite a bit to unpack in this game. Yes the Saints absolutely throttled the Eagles in their regular season matchup, but the Eagles are certainly a much different team this time around and they come into this game with a few things that could work their way.

The Eagles are of course coming off a win against the top-ranked defense of the Chicago bears. In particular, they’ll be able to take what helped beat the top-ranked rush defense in the conference and apply that to the Saints rush defense which trailed the Bears’ unit by one spot. In addition, the Saints defense possesses a weakness against the running back position in the passing game as they have the fourth-worst pass defense DVOA against RBs. That means ex-Saint Darren Sproles may find himself in the mix quite a bit this time around, as he was yet to return from his near season-long injury by the time of the first matchup.

Establishing their running backs as a receiving threat will help keep the Saints defense modest, and should help open up the deep ball which has killed the Saints this season who rank as the worst in the league against passes of 15 yards or more. The problem is that Nick Foles has not always been a great candidate to throw deep. Last week Foles completed just two of his seven attempts at that range, including an interception (although he did have two 14 yard completions, one of which went for a TD). Whether the Eagles will get the Foles who averaged 7.2 and 5.3 yards per attempt in the last two regular seasons or the Foles that averaged 9.2 YPA in the 2017 playoffs will probably determine how successful the Eagles’ passing attack will be come Sunday night.

As for the model, the Saints are made to be three point favorites against the Eagles on a neutral field when coming off a bye. The remaining game-specific factors then bring the spread up to NO -7.1, which is just 0.9 points off the current line of NO -8. That of course means that the Saints would become a model play beginning at -5, whereas the Eagles would become one at +9.5 (but the model removes spreads of +7 or higher from consideration).

That will wrap it up for the Divisional round, but I will see you all next weekend for the conference championships. Don’t forget to follow me on Twitter, where you can catch model plays the second they are made.

  • 1/4: Saturday Wild Card
  • 1/5: Sunday Wild Card
  • 1/11: Saturday Divisional
  • 1/12: Sunday Divisional
  • 1/19: Conference Championships
  • 2/2: Super Bowl

Thanks again for reading!


NFL Playoffs: Division Round Saturday

What a great way to start the playoffs! A nice 2-0-1 Wild Card weekend for +4.6 units also brings highlighted newsletter plays to 5-1-1 for +15.2 units. As a whole, the model moves to 74-43-6 for +93.69 units and a 20.18% ROI. The Wild Card round is always intriguing since division winners get home field advantage over Wild Card teams, creating situations where sometimes the better team is forced to play on the road which then in turn creates super close coin-flip matchups. The Divisional round is also interesting, as every year talking heads and fans alike weigh in on the debate of whether a bye or the momentum from winning a Wild Card game is more valuable.

The Value of a First Round Bye

Teams with a first-round bye are 44-21 in divisional round games since 2002, when the current playoff format was adopted. That’s good for a 67.7% win percentage for those teams, which is pretty significant at first glance. But how do we discern how much that advantage is from playing at home and how much is from coming off a bye? How do we separate that distinction from the fact that teams receiving byes are generally stronger teams to begin with and should be winning games at a higher clip? Looking at just playoff splits can be a bit deceiving due to small sample size, as a 65 game sample across 15 years is laughably small. As a result, I’ll look at splits using regular season games to capture a larger sample.

Since 2002, home teams have won 57.6% of games in the regular season and teams coming off of a bye (home or away) have won 55.2% of their games. If we look at a split combining the two, home teams coming off a bye have a 62.8% win percentage. Given that most first round bye teams are favorites for their Divisional round game, I thought it’d be worth mentioning that home favorites have a 73.9% win percentage coming off a bye since 2002. But these are all just in terms of wins and losses. The question for our purposes,should shift to how effectively Vegas has accounted for bye week value. Home teams off a bye are 53.6% ATS; home favorites off a bye are 55.0% ATS. Although those splits are “winning”, Vegas has been able to take advantage of bettors trying to take advantage of the bye week angle as of late. Since 2010, home teams off of a bye are 43.3% ATS and were 3-10-1 this year alone. With this in mind, it is very likely that Vegas has in recent years intentionally baked in more value than necessary on teams coming off a bye week in order to capitalize on public bettors trying to take advantage of the angle.

In my opinion, I think the real-world value of a playoff bye is from just being able to play less games in order to reach and win a Super Bowl. Without a bye, winning four straight games against above average to elite teams is a big mountain to climb considering winning four straight games in the NFL is tough enough as-is. Reducing that to three games with at least one game guaranteed to be at home and another on a neutral site is a much more manageable task. But this is all a digression and the matter of the fact is that I do account for the additional edge of a team coming off of a first round bye in the playoffs, but it is likely not as large of an edge as most bettors make it out to be.

#6 Colts (11-6) vs. #1 Chiefs (12-4)

Last week I called T.Y. Hilton the key to the Colts-Texans game, and surely enough he delivered. A nice 5 rec / 85 yard game helped extend key Colts drives while also spreading the field to make everyone else’s job on offense easier. Eric Ebron, the other candidate I called out as a potential producer, also contributed by securing a touchdown. Marlon Mack also joined in on the fun, tearing apart the highest-rated rush defense for 148 yards and a touchdown. But enough about last week, let’s get into the Colts’ matchup this week against the awaiting top-seeded Kansas City Chiefs.

The Chiefs’ journey in the model has been a very interesting one, to say the least. Prior to the season I had them Chiefs as the ninth-best team overall and fifth-best AFC team with 9.10 expected wins, and projected them as a fifth seed for the playoffs. My “low” evaluation didn’t come from me not being a believer in Patrick Mahomes, as I had the Chiefs improving their pass offense up to fourth-best heading into the season. Obviously what he and that offense has been able to achieve this season has blown anyone’s projections out of the water, and it should come as no surprise that the Chiefs’ pass offense ranks first by quite the margin. In fact, the gap between them and the second-ranked Chargers’ pass offense is as large as the gap between the Chargers and the twelfth-ranked Eagles unit.

One thing I have talked about on multiple occasions is the model’s ability to identify incorrect evaluations and adapt. Despite starting the season as the ninth-best team, the model had already jumped the Chiefs to the #2 spot after just three games and improved their expected wins total by 1.71, which is a massive jump for that time frame. The Chiefs’ peak this season came after Week 10, at which point they had 14.27 expected wins – a mark that no other team has come close to touching this season. The Chiefs of today aren’t as impressive in the model as they were at their peak, but they still rank first with 11.75 expected wins. They also possess the third-best improvement in expected wins over the course of the season, a mark that the Colts coincidentally bested as covered in last week’s newsletter.

As for this week’s matchup, I think we will see Andy Reid rely a lot on 12 personnel (two tight end sets). The Chiefs ranked third in the league in 12 personnel play frequency (26%) and the Colts ranked third-worst in defensive success rate against that package (59%) while allowing a 112.0 passer rating in those situations. They also surrender the most yards per game to tight ends and rank 29th in pass defense DVOA to the position. In last week’s newsletter I mentioned All-Pro rookie linebacker Darius Leonard who is certainly the anchor of the Colts’ front seven, but his ability (PFF coverage grade: 78.9) has not been enough to hide Anthony Walker (55.7), Matthew Adams (41.6), and Zaire Franklin (41.4) when they are called upon to cover two tight end sets. With this in mind, second-string Chiefs tight end Demetrius Harris is perfectly poised to have a career performance and doing so may help his agent fool a team into paying him the big bucks this offseason.

As for the model, the Chiefs are made to be 6.2 point favorites. The current line shows KC -5, which isn’t a large enough disagreement to make a model play on either side. If the line were to move to KC -4, then the Chiefs would become a model play at 2.2 units with each additional half point on the spread away from 6.2 being another half unit added to the play. The Colts would in theory become a model play at +8.5 or higher, but the model excludes spreads of +7 or higher from consideration and I have yet to make a decision on whether that exclusion should be lifted for the playoffs and will put off that decision until need be.

#4 Cowboys (11-6) vs. #2 Rams (13-3)

The Rams are kind of an unexciting team to discuss as they’re one that many pegged as being a top team heading into the season and they’ve of course finished as such. The model is no different, as the Rams entered the season as the #2 with 10.69 expected wins and finished the regular season as the third-best team with 10.53 expected wins. That may seem low given that the Rams finished with 13 wins, but their adjusted Pythagorean win total was 11.15 for this year. The one area in which the Rams did surprise many was their awful run defense, which finished 28th in the model despite the presence of Aaron Donald and Ndamukong Suh up front. This soft spot will certainly be one that Jason Garrett can send Ezekiel Elliott at, who just tore the Seahawks up for 127 yards on 16 carries (7.94 yds/att).

Outside of that, Jason Garrett will have to come up with a gameplan in the passing game that vastly differs than the one executed last Saturday, as a lot of what yielded success for them in that game will be unlikely to bear fruit again this Saturday. Of Dak Prescott’s 33 attempts, 27 of them were for 15 yards or less and only eight of those were thrown to the left of the left hash mark. If this trend continues, the Rams are going to absolutely feast on Dak. The Rams defense ranks sixth overall in DVOA against short passes as a whole, and ranks second in pass DVOA against short passes (15 yards or less) to the right. Going the complete opposite – deep and to the left and/or down the middle – would be much more beneficial for the Cowboys offense, as the Rams defense ranks 23rd and 21st in those directions. Given that Prescott has only attempted 8% of his total passes this year to deep left and deep middle and a whopping 85% 15 yards or shorter, the Cowboys offense might find themselves in a lot of trouble early and often.

What about the Rams? Concerns on the offensive side of the ball started to surface after putting up six points against the Bears and 23 against the ravaged Eagles secondary in back-to-back weeks. Those concerns quickly went away as the Rams seemingly returned to form, putting up 31 and 48 in the final two weeks. Sean McVay has received endless praise this season for the success of the Rams’ offense, who has kept it simple and have run 96% of their plays from the same formation. Although McVay is certainly a great coach, I think the praise has been a bit exaggerated. If you exclude their Week 1 cupcake matchup against the Raiders, the Rams have faced the eight-toughest schedule of pass defenses. That may make their offensive performance this year look even more impressive, but if you look at the Rams’ schedule from Week 7 onwards, the Rams have faced eighth-easiest schedule of pass defenses. The problem is that the Cowboys don’t exactly possess the talent to adequately challenge the Rams, as they rank league-average in pass defense anyways.

As for the model, the Rams are made -5.5 favorites for this game. With the Vegas spread set currently at LAR -7, there is no model play currently but the Rams would become a play starting at -3.5 (don’t expect to see this line) and the Cowboys would technically become a play at +7.5, which comes with the same caveat the Colts did in the other game’s write-up. If potential model plays on IND or DAL show, I will be sure to make a decision and communicate it via Twitter. As of now I am siding with not lifting the exception in order to maintain the status quo, but I admittedly can’t think of any other reason not to. We shall see.

That’s going to wrap it up for today. Don’t forget to check your inboxes again tomorrow for a write-up covering the Sunday games. As per my Twitter, there is already a locked in model play on LAC +4.5 which I will be covering in-depth for that write-up. Below you can find a schedule for the remaining NFL Playoffs write-up schedule.

  • 1/4: Saturday Wild Card
  • 1/5: Sunday Wild Card
  • 1/11: Saturday Divisional
  • 1/12: Sunday Divisional
  • 1/19: Conference Championships
  • 2/2: Super Bowl

See you soon.


NFL Playoffs: Wild Card Sunday

Today we will be finishing out the Wild Card round by taking a look at the two Sunday games. To begin, lets take a look at the noon matchup.

#5 Chargers (12-4) @ #4 Ravens (10-6)

Every NFL season has at least one or two teams that make me go, “Wow, that team definitely deserved to be in the playoffs”. This game happens to have the two teams I thought were most undeserving of missing out on last year’s playoffs. For those who don’t remember, the Chargers started 0-4 last year which was a stretch that was full of incredibly unlucky circumstances (and that has been the Chargers’ “thing” for quite some time now). They then went on to go 9-3 for the remainder of the season, with their losses coming all on the road against eventual playoff teams (Patriots, Jaguars, and Chiefs). They finished with 10.46 adjusted Pythagorean wins, good for fifth in the conference and ahead of the Chiefs (9.99), Titans (7.38), and Bills (6.35) – all of who made the playoffs last year. The Chargers differential between actual wins and adjusted Pythagorean wins (-1.46) was the sixth largest in the league in 2017, which was a sign that they were potentially due to put together a better campaign in 2018.

The model seemed to agree. Heading into the 2018 season, the Chargers were the highest ranked team in the model, ranking first in pass offense and third in pass defense. The Chargers ended up finishing second overall in the model with 10.53 expected wins and within 0.35 expected wins of their preseason projection. They also finished top ten in every model category except special teams, finishing second in pass offense, sixth in rush offense, tenth in pass defense, and tenth in rush defense. On top of their excellent finishes, the Chargers have shown incredible consistency in the model having never ranked lower than third in the model at any point this season. Some of you may be asking, “If the Chargers are so good according to the model, why did they get hosed by the Ravens in Week 16?”. It’s a valid question, for sure.

In my Week 17 write-up, I highlighted the Ravens’ ability to grind clock through their relentless running which in turn vastly reduces the amount of offensive opportunities opposing offense have. The Chargers actually did not fall victim to that trend in the Week 16 matchup as they ran only two fewer plays than their season average and had 28:35 in time of possession to the Ravens’ 31:25. And it wasn’t Lamar Jackson’s running that decidedly beat the Chargers – he turned in a career-low 39 rushing yards while passing for a career-high 204 yards. So how did the Ravens pull of the Week 16 upset?

Although a 22-10 score may tell a different story at surface level, this game was very close to going the Chargers way. With 5:29 remaining in the 4th quarter, the Chargers were down 10-16 and had a 3rd and 5 on the Ravens’ 29. Philip Rivers on this drive had already converted three straight third downs and was putting together the Chargers’ largest drive of the game. He was then sacked for an 11 yard loss, pushing the Chargers out of field goal range. Luckily, they were able to pin the Ravens on their own 2 on the ensuing punt and held the Ravens’ offense to a three-and-out. The Chargers got the ball back on the Ravens 39 yard line and just 39 yards sat between the Chargers and a likely number one seed for the playoffs. On first down, Melvin Gordon ripped off an eight yard run which was called back by holding. Then on the next play, Antonio Gates fumbled the ball which the Ravens returned for a 62 yard touchdown.

On top of that ending sequence to the game, this game had a lot of unfortunate and uncharacteristic things go against the Chargers. On offense, their first offensive play led to an interception and then they also had three early third down conversions negated by penalties which killed drives. On defense, the Chargers allowed a 68 yard touchdown to rookie tight end Mark Andrews. This was particularly egregious given that the Chargers are the top-ranked defense against tight ends according to DVOA. These kind of things are not the type that I would expect to happen regularly. Conversely, the Chargers are likely the most injured team coming into the playoffs and that could most certainly work to their disadvantage for this go-around.

Ultimately, this week’s rematch will likely come down to who commands the lead when the fourth quarter begins. The Ravens are the only playoff team to not have a comeback win when trailing in the fourth quarter. And although Lamar Jackson has a superb 6-1 record to start his career, every one of his starts have featured a one-score game in the fourth quarter. That becomes a bit scarier when you consider some of the teams the Ravens have played during that stretch (CIN, OAK, ATL, and TB). It would be wise of the Ravens to try to get a lead early on the west coast Chargers who will be travelling across the country for an early game in which their body clocks will be set to 10 AM. Either way, this game is certainly my favorite of the opening round. On one hand we have a very experienced quarterback with a coach making his playoff debut going up against the youngest starting playoff quarterback ever with an experienced and successful playoff coach (Harbaugh is 10-5 all-time in the playoffs).

The model gives the Chargers a 53.64% chance of winning this game on a neutral field, which would be good for a LAC -2 spread in such scenario. Even though the Chargers don’t really have a “home field” and essentially have played every game of this season as the away team (and are 7-1 in actual away games this season), the model does make this game BAL -0.9 after those adjustments. That means that LAC +3 -110 is a model play risking 2.300 units to win 2.1 units. That makes it the first official model play of the NFL playoffs. Exciting stuff.

#6 Eagles (9-7) @ #3 Bears (12-4)

In yesterday’s newsletter, I mentioned that the Dallas Cowboys were the most undeserving team to make the playoffs based on expected wins given their 23rd rank in that regard. The Philadelphia Eagles would be the second-most undeserving team to make the playoffs, as they rank as just an average team in the model at 15th with 8.26 expected wins for the season. Granted, the Eagles do rank near the top ten in pass offense which is and has been the most important model category of the season by a wide margin. The problem is that the Bears rank first overall in pass defense, and given that the Eagles have one of the league’s worst rushing attacks and are going up against the second-best rush defense, the Eagles will certainly have to try to make the passing game work in whatever ways they can. But they will have to do so on early downs as the Bears have the best third-and-long defense ever recorded according to Football Outsider’s DVOA metric.

The good news is that although this match up certainly favors the Bears, the task may not be as large as Vegas is making it out to be. The Bears themselves aren’t too far off in the model from where the Eagles stand, as they currently sit as the tenth best team with 8.90 expected wins. This may seem odd to some given the fact that they have twelve wins on the season and given how absolutely dominant their defense has been. A lot of that actually lies on the strength of their opponents, as the Bears have faced the easiest schedule of opposing defenses and the thirteenth-easiest schedule of opposing offenses according to Sharp Football Stats.

But as some of you may point out, I have already shared that the Bears have the top-ranked pass defense and second-best rush defense in my model, which adjusts for opponent strength. So why aren’t they higher in the model? For those of you that have joined the Bet It Up newsletter recently, my initial introduction to the model highlighted that the model dynamically weighs five categories (pass offense, rush offense, pass defense, run defense, and special teams) based on their calculated correlation to team strength across the league. Those weightings are exactly why the Bears may not be as terrifying as their defense is. Pass defense and rush defense combined are not weighted as much as even just rush offense alone. And pass offense, if you’ve been following along, has significantly more weight than rush offense.

It’s because of those weightings the Bears’ strength in the model and the dominance of their defense is certainly dragged down by their below average pass and rush offenses. And that may be exactly why the model makes this game CHI -4.2 which is a much shorter spread than the current spread of CHI -6.5. Short enough that PHI +6.5 -104 qualifies as a model play risking 2.403 units to win 2.3 units.
That will wrap it up for today’s write-up as well as this week’s round of playoff games. Don’t forget to tune in for next week’s matchups as well as the remainder of the playoffs (write-up schedule below). As always, you can catch model plays the second they’re made by following me on Twitter.

  • 1/4: Saturday Wild Card
  • 1/5: Sunday Wild Card
  • 1/11: Saturday Divisional
  • 1/12: Sunday Divisional
  • 1/19: Conference Championships
  • 2/2: Super Bowl

Until next time!

NFL Playoffs: Wild Card Saturday

Last week’s highlighted newsletter play of CLE +6 pulled through for the readers here bringing highlighted plays to 3-1-1 for +10.60 units, but the rest of the model’s Week 17 plays weren’t as hot. The week finished 2-2 for -4.51 units, bringing the model’s record for the regular season to 72-43-5 for +89.27u and a 19.51% ROI. With the regular season wrapped up, it’s time to shift our focus to the playoffs. I will be covering every playoff game for the readers here, providing the model’s spread as well as analytical insight on the matchup and the underlying model data. To start, lets take a look at the Saturday set of Wild Card games.

#6 Colts (10-6) @ #3 Texans (11-5)

The first playoff game of the season will pit two teams that the model would have never predicted to be in this position before the season began. The Colts’ 2017 season was of course hampered by the entire Andrew Luck shoulder fiasco, but a lot of the team’s performance despite that was very concerning. The offensive line allowed a league-worst 56 sacks and on the other side of the ball the Indianapolis secondary ranked dead last against the pass. The Luck shoulder fiasco then bled into the 2018 offseason, camp, and preseason as concerns of his ability to throw at even moderate lengths were in question. Heading into the season, the Colts ranked 27th in the model and were only expected to win 6.49 games. After a very flat and uninspiring 1-5 start, the Colts looked to be trending to be exactly who the model thought they were.

What Frank Reich and the talent on that team has been able to do since then has been nothing short of impressive. In their 1-5 start, the Colts averaged 25.3 points scored and 30.0 points allowed. Since then, the Colts have averaged just 16.4 points allowed and have averaged 31.2 points scored if you remove their Week 10 6-0 dud against the Jaguars. A lot of the team’s success has come on the back of some new faces. Sixth overall pick RG Quenton Nelson not only finished as the highest-rated rookie offensive lineman by Pro Football Focus, he also finished sixth at the position overall. Nelson’s performance helped turn the league-worst offensive line unit into the league-best, with the Colts finishing with the least amount of sacks allowed (18) as well as finishing second in adjusted sack rate. On defense, second-round pick Darius Leonard also turned in a Pro Bowl-caliber season (but was unfortunately snubbed of the honor), and finished second for rookie linebackers and sixth overall at his position. Interesting enough the model caught on to the Colts’ turnaround pretty quickly, upgrading them following their Week 8 win against the Raiders to 8.42 expected wins, which ranked 15th at the time. The Colts have since climbed even further, finishing the regular season ranked ninth in expected wins with 9.07 expected wins. The Colts’ expected wins increase represents the second-largest in the model for this season.

The Texans have had a similar trajectory as the Colts, having started the season as the 28th ranked team in the model and only expected to win 6.37 games. The Texans had a poor start as well, dropping its first three games. However, the model actually saw in the underlying data that the 0-3 start was somewhat deceiving and even bumped them to 7.16 expected wins after that stretch. That 0.79 increase in expected wins intrigued me as the ensuing nine game win streak that followed led to a 1.43 expected win increase. In other words, the Texans’ performance during their nine game win streak was just short of being twice as impressive to the model as their performance during their 0-3 start. Today the Texans sit in the model as a team expected to have won 8.74 games, good for 12th best in the league. Their expected wins increase represents the fourth-largest in the model this season.

As for their matchup on Saturday, you can infer from their close current expected wins numbers that these two teams are matched pretty evenly. The Colts are given a 50.9% chance of winning on a neutral field, good for a IND -0.5 spread. After including game-specific adjustments, the model’s line for the game is made to be HOU -2.6. The Vegas line as of the time of this writing is HOU -1.5, and at that line there is no suggested model play since model plays are when the disagreement between the model line and Vegas line is by two or more points. However, there would be value on the Texans starting at HOU -0.5 (2.1 units) and there would be value on IND starting at IND +5 (2.4 units). Obviously each half point further from each of those spreads would then be matched with an additional half unit on the play (example: HOU +1 would be a 3.6 unit play).

As for the game itself, I believe the key to this game will be Colts wide receiver T.Y. Hilton. Hilton has been on an absolute heater as of late: in the last seven weeks he has led the league in receiving with 840 yards (or 120 yards per game), which is 163 yards more than the next best receiver during that span and all coming despite battling an ankle injury. Secondly, he has torched the Texans’ secondary for quite a while now. In his last five games against the Texans, T.Y. Hilton has averaged a 6 rec / 107.6 yds / 0.6 TD line which includes a 9 rec / 199 yds line in the most recent meeting. Hilton burns the Texans’ secondary for good reason: Houston ranks 29th in yards allowed to WR1s and they rank 31st in DVOA pass defense against WR1s. Eric Ebron is another candidate to tear the Texans’ pass defense apart given that they rank 31st in yards allowed to tight ends and 23rd in DVOA pass defense against the position. In his two games against HOU this year, Ebron has accumulated 9 rec / 105 yds / 2 TDs.

#5 Seahawks (10-6) @ #4 Cowboys (10-6)

Saturday’s NFC Wild Card game will feature another team that has greatly outperformed the model’s preseason expectations: the Seattle Seahawks. I already covered their journey in some depth in my Week 14 write-up, and for those of you who weren’t subscribed to Bet It Up back then (shame on you), here is what I said then:

“The Seahawks on the other hand were a very unimpressive 9-7 team in 2017 that performed well above their 6.57 adjusted Pythagorean expected win total. The team then proceeded to lose Richard Sherman, Michael Bennett, Paul Richardson, Sheldon Richardson, Jimmy Graham, and more in the offseason, and were set to be without their top wideout Doug Baldwin for an indeterminate amount of time. They also burned their first round pick on San Diego State running back Rashaad Penny, which I thought was an awful use of the pick given that they already had the serviceable Chris Carson and had more glaring team needs. With all of this in mind, I’m not ashamed to share that by the end of the preseason the Seahawks were the 29th best team in the model, expected to win only 6.31 games in 2018.

I then went on to share that the Seahawks had gradually improved up to sixth in the model ahead of their Week 14 game against the Vikings. The Seahawks ended up finishing sixth with 9.50 expected wins, which was a 3.19 expected wins improvement – the largest of any team in the model this year. This is largely in part due to the well-roundedness of their team, with the only model category they failed to land inside the top ten in being run defense. A lightbulb may suddenly go off in your head as you envision league-leading rusher and Cowboys franchise running back Ezekiel Elliott on the other side of this contest. However, Zeke has had somewhat of a deceiving season. From Football Outsiders Quick Reads: 2018 in Review article, “Elliott led the league in rushing, but was just ninth in DYAR and 18th in both DVOA and success rate. He led all running backs with six fumbles on running plays — or, one for each rushing touchdown he scored”. Elliott’s more meaningful contributions may come in the passing game, as the Seahawks rank 26th in running back receiving yards allowed. Or potentially the threat of Elliott running could force the Seahawks to stack the box and allow trade deadline acquisition Amari Cooper or rookie Michael Gallup to exploit the Seahawk’s 28th-ranked defense against explosive passing (courtesy of Sharp Football Stats).

Either way, the Cowboys will have to find some avenue of success to exploit early and often as they are by far the most undeserving team to make the playoffs according to the model. The model has them as a 7.539 expected wins team, which puts them 23rd overall. With this in mind the model makes this game on a neutral field SEA -2.5 but home field factors bring this game to a near pick ’em, with the model spread being DAL +0.1. The Vegas spread for this game at the time of this writing is DAL -2, which would qualify SEA +2 as a model play for 2.1 units. However, I am anticipating a move to at least SEA +2.5 juiced at -105 or lower. As a result, I will not make the play on SEA +2 an official model play as of right now but I will be tweeting out the model play when I do officially make it. If by some magic the spread hits DAL +2.5, that would become a model play on DAL for 2.4 units.

That’s going to wrap it up for today’s write-up. I hope you guys enjoyed it, and don’t forget to check your inboxes for another write-up tomorrow as I will be covering the Sunday Wild Card games as well. As a reminder, here is a schedule for my playoff write-ups:

  • 1/4: Saturday Wild Card
  • 1/5: Sunday Wild Card
  • 1/11: Saturday Divisional
  • 1/12: Sunday Divisional
  • 1/19: Conference Championships
  • 2/2: Super Bowl

Thanks again for reading!

NFL Week 17: Mayfield’s Motivation

Last week was admittedly not the brightest of spots for the model, going 1-3-1 for -8.495u and generating the first losing week since Week 9. A performance like that is bound to pop up every so often given the size a one week sample is in the grand scheme of things. The highlighted newsletter play of HOU +2 -104 (2.403 for 2.3 units) pushed, with Nick Foles turning in his best regular season performances by most metrics since 2015. Highlighted plays move to 2-1-1 for +8.50u since the model’s introduction to the readers here and the model overall moves to 70-41-5 for +93.78u and a 21.15% ROI. Week 17 (luckily) has a lot at stake for a handful of teams, and should make for an entertaining week of football.

As I’ve mentioned before, spreads of +7 or higher have been removed from model consideration since Week 9. With so many of those spreads on the board this week, there are slim pickings when it comes to playable games. Luckily for us, there is one model play that is not only worth making, but one that includes two teams that I’ve been itching to talk about in detail for a while now.

There may not be two hotter teams in the NFL than the Browns and the Ravens, both 5-1 in their last six games. The Ravens’ run has come on the back of their incredible defense and the change at quarterback to Lamar Jackson, who with the help of John Harbaugh has completely changed the offensive identity of the now AFC East-leading Baltimore Ravens. A win on Sunday would secure an AFC East title and a four seed in the AFC, likely pairing them in a rematch with the Chargers (or a less-than-likely rematch with the Chiefs) during the Wild Card round. A loss would have them watching the Wild Card games from home instead.

As for the Browns, their run has come on the back of firing Hue Jackson. A win on Sunday doesn’t do much for the Browns as they are mathematically eliminated from playoff contention, very largely in part due to Hue’s incompetence which costed them potential wins earlier in the season. The common angle public bettors take year-after-year in Week 17 is that teams with something to play for (like the Ravens) should easily win and cover against teams that have nothing to play for (like the Browns). But do the Browns truly have “nothing” to play for?


Cleveland Browns quarterback Baker Mayfield has not been shy when it comes to sharing what he believes he and the rest of team is capable of (and presumably indirectly, what Hue Jackson prevented them from accomplishing). This week is no different, as Baker Mayfield made it very clear what’s at stake for both teams: “They’re fighting for a playoff spot, and we’re fighting to prove who we are” (via ESPN). So now that we can clear this game of any motivational questions, we can take a higher-level look at the matchup.

The Ravens finished the 2017 season with a 9-7 record and without a playoff berth. Joe Flacco’s play and contract continued to draw criticism from Ravens fans and the front office made an attempt to give Flacco a better chance at succeeding (signing Willie Snead and Michael Crabtree, drafting Mark Andrews and Hayden Hurst) while also warming up his seat a bit (drafting Lamar Jackson) to signal the urgency to perform in 2018.

Everything coming out of camp from Ravens beat writers signaled an invigorated Flacco and Baltimore offense, and pairing that with an already elite defense convinced me enough to project Baltimore in the model as the fifth-best team heading into the season. They would hang around that range during their 3-1 start until the Ravens stumbled through a 1-4 stretch during which Flacco would sustain a hip injury, and Lamar Jackson would be named the starter following their Week 10 bye.

Even before Flacco’s injury, you can see above that from Week 7 to Week 9 the Ravens pass offense had already began to decline whereas its run game began to bounce back. The shift from Flacco to the run-first, pass-shy Jackson gave John Harbaugh a chance to expand on these trends and completely shift the identity of the offense. The change to a run-first-and-keep-running offense also played into the Ravens ultimate strength: its defense. Running the ball at absurd rates meant the clock was moving more on offense compared to the previous iteration of the Flacco-ran offense, which works twofold in their favor. Opposing offenses are having to do more on offense with less opportunities. Per Evan Silva, teams are averaging 10.9 less offensive snaps per game against the Lamar Jackson Ravens compared to their season average. That is a 17% decrease in offensive snaps on average, meaning teams are losing almost a fifth of their offensive opportunities. Secondly, less opposing offensive snaps obviously means less defensive snaps for the Baltimore defense, which keeps their defense fresher and sharper which makes defending those limited opportunities easier. The spike in pass defense above illustrates this rather clearly.

Now despite this massive change and a 5-1 run, the Ravens’ expected wins number has only moved by 0.028 since the change. But sometimes staying afloat while others unravel is all you need in this league to survive. Before naming Lamar Jackson the starter, the Steelers were a surefire bet to win the AFC East and in the model there were four teams with two or more expected wins than the Ravens. Today, the Ravens lead the division and in the model there is only one team with two or more expected wins than the Ravens.

Shifting to Cleveland, the Browns were a team during my offseason deep dive and projection process that had a lot of talent I loved. Jarvis Landry was a great talent to add, Josh Gordon was seemingly going to turn it around (again), David Njoku was a developing freak at tight end, Nick Chubb was a draft prospect that really popped off the page, and Myles Garrett and Denzel Ward were incredible young defensive talents who I trusted to anchor their respective segments on defense. On the other hand, I admittedly was not the highest on Baker Mayfield and thought he would be above average at best. I did think Mayfield was better than then-starter Tyrod Taylor, but I also expected Hue Jackson to botch the handling of that situation (as well as the rest of the team). Two years of laughably bad coaching was enough to convince me that this team was doomed as long as that man was in town, and I projected them as the third worst team in the model.

The Browns season has been covered in much detail, so I don’t feel the need to recap how their season has gone. The above chart is a look at how the team has fared before and after Baker Mayfield becoming the starter and Hue Jackson getting the boot. Note that Cleveland’s pass offense under Hue Jackson peaked in the Jets game in Week 3 when Mayfield had to come in for the injured Tyrod Taylor.Also note that the pass offense got progressively worse from that point on until Jackson’s firing. Then note the absolutely incredible improvement to the Browns’ pass offense following his firing. Hue Jackson is a cancer to every talent, team, and organization he finds himself meddling with and I am absolutely bewildered that he was able to find legitimate work in the NFL so quickly (but am not surprised it was with the Bengals, of all teams).

For their Week 17 matchup, the Ravens and Browns currently rank 9th (9.000) and 16th (8.018) in expected wins in the model. This gives the Ravens a 52.88% chance of winning on a neutral field, good for a -1.5 spread under those conditions. After factoring in the various game-specific elements the model makes this game BAL -4.4. With the line set at BAL -6.5, there are 2.1 points of disagreement between the model and Vegas lines which makes CLE +6.5 -105 a model play risking 2.215 units to win 2.1 units.