FanPost

To Invest or Not Invest... That is the Question

I know that we are in Patriots week, in the middle of the Holidays, and the Bills are in the thick of a playoff run, but I wanted to share some data that I have been tweaking off and on for a couple of years. Most of it I have shared before, but I have updated it and accumulated the data for 3 years (2019, 2020, and 2021). I think the result of this analysis could help us frame the current Bills' season and it can also help us focus in on what will help us in next year's offseason. You can be the judge of that though!

To that point, there has been lots of talk about what is wrong with our running game. Is it the RBs? Is it the OL? And there has also been some angst over how the Bills have easily dismantled several bad teams, but have struggled against some good ones. All of this has prompted more discussion on whether the Bills should draft this position or that position this offseason. Is it better to draft a skill position player or a guy in the trenches?

To help shed some light on these questions I have tried to gather some team performance metrics that can be correlated against offensive positional "investment values". This can hopefully tell us if investment in one position or another is generally more important than the other positions.

Offensive and Defensive Performance Metrics:

A month or so ago, I shared some data on some of my favorite Offense and Defensive stats. https://www.buffalorumblings.com/2021/11/3/22762561/offensive-and-defensive-rankings?_ga=2.80329502.867033031.1639777373-551707090.1584535299

The Post was based on 4 offensive/defensive metrics: total yards/play, points/drive, net pass yds/att, and rush yds/att (for RBs only). While these are not perfect stats (no stats are!), this collection of metrics does convey the average efficiency of an offense/defense. In the post above, the main concept was that the performance metrics were not just standard off-the-shelf metrics. These metrics were adjusted based on the performance of the team's opponent. Since then, I have updated it to do a 3rd order adjustment. The adjustment is similar to Football Outsiders' DVOA (Defense Value Over Average) adjustment. The difference is that DVOA adjusts the metric based on the NFL average while my adjustment adjusts the metric for only the opponents that each team plays. I call this adjustment OAV (Opponent Adjusted Value). To me, this makes much more sense than DVOA. Teams don't play the whole NFL, so why would you average the metrics based on the whole NFL. Some teams play much easier/much tougher schedules than other teams, so their performance should be impacted by this reality.

Positional Investment Value

I also did a post on Positional Investment Value. https://www.buffalorumblings.com/2021/2/18/22289938/offensive-supporting-talent-position-value?_ga=2.80329502.867033031.1639777373-551707090.1584535299

However, since then I have updated the Positional Value to span three seasons: 2019, 2020, and 2021. This should hopefully smooth out any anomalies in the data. Basically, the system described in the post comes up with a rating of how much a team invested in the player. For players on their rookie contracts, the investment is driven by the draft round that they were taken. For veteran players, it is driven by their contract salary compared to others for their position. Using these two rating scales, we can define the investment value of the player on the day that they make the roster. However, as you know players get hurt and hold out, so even though the team invested in the player they do not always realize all of that investment during a season. So, in order to effectively compare this investment rating to offensive success, I came up with the realized investment of a player/position by multiplying each player's investment value by the player's % of participation for the season. For the OL, I used "games started" to determine their % of participation. For RBs, TEs, and WRs, I used pass targets and/or rush attempts.

Using these two metrics I can compare them against each other to see if there are some reasonable correlations.

An Example of the Correlation:

This chart below just shows an example of correlating the Offensive Line Position Group Value against Adjusted Total Yds/Play. Elite Dual Threat QBs and "Learning QBs" (only 1 or 2 years of starting experience) are called out separately. I removed the teams with "Learning QBs" from the data because I found this group to be anomalies to the data. As you can see "learning QBs" ... well... are learning, the talent around them can definitely have an impact, but even the teams with good ones are mostly below the NFL average. I also thought it was interesting to see how Elite Dual-Threat QBs can be a big advantage in building a potent offense. For the rest of the teams, despite some expected variation there definitely seems to be a trend based on the investment in supporting talent around them (green dotted line). There are very few Vet QB teams below the NFL average (5.3 yds/play) on the right-hand side and there are very few Vet QB teams over the NFL average on the left-hand side of the chart (unless the QB is an Elite Dual-Threat QB).

51764988307_efb126a51e_o.0.png

REMEMBER: This rating is a "rough estimate" based on average trends. It is by no means perfect. As you can see in this chart and as we all know from years of watching sports, many players outplay their contract/draft position and visa versa. Schemes, scheme fit, effort/hard/work, and coaching are also important. That said, I do think the trends tell a story at a macro level about how teams invest in position groups and how that investment translates on average to performance.

The Results:

Here are the summed-up results for the on average impact of investment in each offensive position group for 2019, 2020, and 2021.

51765009537_9d5c7de7bf_o.0.png

Some interesting points...

  • It should be no surprise that "on average" the QBs had the highest impact on performance. However, it may be a surprise to some that the QB was not a bigger percentage. We have been force fed since we could walk and talk that the QBs are 100% responsible for wins and championships. Passing stats are QB stats. They get all of the blame and all of the glory. While it is true that QBs are the most important positions, they are also highly dependent on the talent around them.
  • Investment in Receivers seems to have 2nd highest return on investment outside of the QB
  • Investment in the OL unit seems to be in general more important in the overall success of the running game. That said, there are 5 O-linemen while you may only have 2 or 3 RBs. As a result, a single dominant RB may have a bigger impact on the running success than an individual O-linemen.
  • It was interesting that all of the impacts were decreased in Pts/Drive. I am guessing that this metric may be more influenced by scheme than all of the other metrics. Creative OCs can make a big difference in the red-zone.

So, how does this apply to the Bills?

Even though this data is showing averages across the NFL, I think there are some insights to the big questions for the Bills for this year and for future years...

Why have the Bills easily dismantled several bad teams, but have struggled against some good ones?

We have all seen the Bills' offense be inconsistent from game to game and half to half. And we have all seen the OL struggle at times this year. If you believe in this analysis, you can see how struggles with any one position group could lead to the whole offense under-achieving or at least being inconsistent. You all know how much I value the 1:1 matchups in a game. Usually "inconsistency" in a unit or player has less to do with inconsistency of the player/unit and more to do with the matchup with the players/unit across from them. In my opinion, this has been very true of the OL this year. Every game this year a different O-linemen has taken turns being the weak link in the OL depending on the matchup across from them. The Bills have played some very high-level defensive linemen this year. The Bills investment in the OL was below average in 2019 and 2020 and is just about at the NFL average in 2021. With that level of investment, you will definitely see some ups but you will also see some downs... or at least need your QB to do some heroics to overcome it. I think that the Bills offense will only go as far as the OL will take it. They don't have an offense that wants to win with short throws. They are getting better at it this year, but their base offense depends on intermediate to deep routes where Josh reads the field from high to low. With this kind of scheme, the OL is more important than with teams that use a short quick-read offense. While I think a good portion of the OL issues is technique and coaching, I do think that this unit needs an infusion of talent too.

How do we fix the run game? Do we need to add a RB or OL?

Regarding the great debate about whether the RB or the OL is the cause of the Bills woes with running the ball, the analysis in this post does not answer that question. I think that film study of every play is the only way to determine that answer. However, this analysis does seem to suggest that in general both matter, but the OL may matter more for the running game... "on average". If you add in the impact that the OL can have on the passing game, then it seems like an easy choice. That said, I think this analysis also provides room for both sides of the debate. While a great OL seems to matter more than a great RB group, the OL unit is 5 players where the RB group is usually 2 or 3 players. So, acquiring a single RB with elite skills may have a bigger impact than a single O-lineman. That is where I am at. I could support upgrades at both position groups. While I personally think the biggest problem with the running game and passing game is the OL based on film analysis, if there is a RB available that has elite difference-making running skills and is an elite pass-catcher, I would be all for acquiring one if they fell to us (in any round of the draft). These kinds of RBs with elite skills are so rare though. Just because a RB is taken in the 1st round does not mean that they are elite. Many 1st round RBs are drafted in the 1st round (or any round) because they were on great college teams with great offensive lines. This makes it very difficult to separate who is the real talent. For the Bills, the key is that if they do add a RB he needs to be a difference maker. We already have some decent RBs that have decent power, decent elusiveness, OK speed, and OK hands. Most RBs in the draft or free agency will be just a change and not necessarily a change for the better. I am not interested in 1-dimensional RBs. Don't give me a small scat back or a big bruising slow back. I don't even want a great running back with stone hands. The bottom line is that I would not just draft an O-lineman or RB high in the draft just to draft an O-linemen or RB. That is what Miami did with Austin Jackson and Pittsburgh did with Najee Harris. I would rather have a 3rd round Spencer Brown-type or an Antonio Gibson/JK Dobbins-type O-lineman or RB than an average 1st round O-linemen or RB. Basically, I want Beane to be smart and not lock into a position or a faux talented player.

I would be interested to see what you see in the data or on these "big questions". I know many of you have very strong opinions on this topic, so I am looking forward to a healthy debate! Let me know in the Comments.


Just another great fan opinion shared on the pages of BuffaloRumblings.com.