• Kansas City Chiefs QB Patrick Mahomes enters the season as the favorite in PFF’s simulation. He is currently tied for the third-best odds in the betting market.
• Justin Herbert and Joe Burrow are the most overpriced quarterbacks on the MVP betting market. They are both great quarterbacks, but it won’t be easy for them to win their respective divisions.
• The most promising long shots include Tua Tagovailoa, Jameis Winston and Trevor Lawrence.
While winning the Lombardi trophy is the ultimate goal of every player in the league, the MVP award also plays a significant role during the NFL season.
The 2022 MVP is one of the below players in 79% of our simulations. Note that PFF odds are odds we computed with our methodology as explained below, while the betting odds are the implied odds (vig not removed) from the current betting line at DraftKings. Since the vig is relatively high for the MVP market at this point of the season, most MVP bets are actually negative value at this point.
It’s 2022, and we might have reached the point at which Patrick Mahomes is underrated, given that he has only the tied-third best MVP odds right now.
We don’t project Mahomes’ Kansas City Chiefs to be the best team in the NFL this year, but the combination of Mahomes and Andy Reid is still expected to be the most efficient. Given the loss of wide receiver Tyreek Hill, much of the credit would need to be given to Mahomes in this case.
Given how much the media has compared Justin Herbert and Joe Burrow, it’s no surprise that these two are the most overpriced quarterbacks on the MVP market. They are both great quarterbacks, but we shouldn’t forget that there are other top passers in the league, and it won’t be easy for these two players to win their respective divisions.
The interesting name in this table is Matt Ryan, a signal-caller who hasn’t looked like an MVP in recent years. However, it’s notable that Ryan has the best supporting cast since 2016 or 2017; he is also projected to have a favorable schedule of opposing defenses and will have one of the easier roads toward a division title. With him, the overarching question is how to weigh the prospect of him simply being washed against the fact that we already know what he is capable of under perfect circumstances.
If everything breaks right
Every quarterback in this group is very unlikely to win the MVP Award. However, there is a 13% chance that everything breaks right for one of them, resulting in a surprise MVP Award like we almost observed three times in a row from 2017 through 2019, with Carson Wentz, Patrick Mahomes and Lamar Jackson, respectively.
So you’re telling me there’s a chance?
All these quarterbacks are 100/1 or longer to win the award, but their collective odds still add up to 5%. The most promising candidates from this group should be Tua Tagovailoa, Jameis Winston and Trevor Lawrence.
Two years ago, we created a model to forecast the MVP race by simulating each quarterback’s end-of-season stats to accurately predict the MVP voting.
Predicting the MVP is pretty straightforward if you know each quarterback’s passing efficiency and win-loss record. Unless there is a quarterback with an overwhelming number of wins (15 or more), the prize usually goes to the most efficient quarterback of the season. Over the last two seasons, this has been Aaron Rodgers.
By forecasting rather than looking at the current situation, we got things right in 2020 when the MVP race was largely between Russell Wilson, Patrick Mahomes and Aaron Rodgers. We also got things right last year when we had Rodgers as a repeat winner when the market still was all over Tom Brady.
Here is a quick reminder of our methodology:
We trained a model to predict the winner of the MVP Award based on only three regular-season parameters after Week 18:
The number of wins and the league-wide rank
Rank in the division
Expected points added (EPA) per play and the league-wide rank
Since the model works with the regular-season stats in hindsight, we need to simulate the remaining games of the season so we can feed the model with the results. We utilize our weekly win total simulations and a Bayesian Updating method to have a prior for each quarterback’s EPA per play in 2022.
The number of wins and the final division standing are obtained from our win totals simulations and win-total simulations derived from market odds. We will draw 10,000 simulations from each and thus simulate the season a total of 20,000 times.
During the season, we update our beliefs on all quarterbacks with Bayesian Updating, incorporating up-to-date EPA numbers and making adjustments based on the PFF passing grade, our accuracy charting and the quality of defense faced.
Using these updated beliefs, we can simulate each quarterback’s EPA per play for the remainder of the season. The new part this year is that we connect the simulated EPA per play and the simulated game results of our simulation on the game level by updating the prior expected EPA numbers with the information of the game result. This also allows us to add dynamic updating to the simulation — the priors of a quarterback change throughout each simulation based on whether he plays well or poorly in this particular simulation.
Here is an example of how this process works. We use the Week 1 Las Vegas Raiders at Los Angeles Chargers game for illustration. Going into this game, the expected distribution of EPA per play for the two quarterbacks looks like this, with Justin Herbert having the higher prior.