10-16-2021, 09:43 PM
(This post was last modified: 10-24-2021, 09:55 AM by allbetsonjames. Edited 1 time in total.)
So this past week I had some fun fiddling around in google sheets trying to create a weighted point system that in aggregate can tell me who the best defensive player is. I used rookies as a small sample size, so please enjoy my Frankenstein rating system for ranking rookie defensive player campaigns this season. I may expand my data set to look at all defenders, depends how bored I am tomorrow. Offense might be a bit of a stretch, maybe next season...
The main gist is instead of looking at raw stat totals, and weighing in my head how impactful this is compared across the board defensively- I created some formulas to spit out the proper leaders with weights baked in. I am taking the % of each stat compared to the total player pool with added weights based upon the frequency that each stat is attributed to a particular position. For example, LB tackles are weighted ~70% while DE tackles are ~190%, and CB TFL are weighted heavily almost 2000%. In order to account for randomness there is a weight based upon totals of the stat recorded. And just to add a cherry on top I added a stat modifier taking into account a simple impact of a stat say compared to a tackle (TFL= x2, FF = x3, Sack= x2, Int =x3, PD = x2)
I started this a week ago and kept just playing with some of the weights/ modifiers to a point where it kind of made sense. So here we are, a look at the top defensive rookies this season and a bit of analysis of the numbers. I’ll start with the overall leaders then we will breakdown each stat category leaders to look at what the system did good and what its weaknesses are. One small note is that I disregarded safeties (since no rookies recorded one) and did not take into account fumble recoveries as they are slightly random.
Here are the formulas that were used to compile these ratings, and the final pizza rating is just the sum of all the individual ratings.
PR=Raw Stat*Volume Modifier*Volume Weight*Stat Weight
where:
Raw stat= normal stat value
Volume modifier=# of players in position ÷ (position stat total*sqrt(positional adjustment))
Positional adjustment= avg stat per player ÷ (position stat total/# of players in position)
Volume weight=total stat #/10
stat weight= (T-x1, TFL= x2, FF = x3, Sack= x2, Int =x3, PD = x2)
Topping off the list we have Teddy Murphy who I think had probably the best all around season of the bunch- he was making plays, disrupting the QB and just an all around stud, a pretty damn good pizza rating of 205 with stellar ratings in tackles, INTs and PDs.
In second we have Murphy's teammate Owen Reed who is actually my front runner for DRoty, he had so much production from the CB position- solid ratings in tackles, amazing TFL representation and a standout with forced fumbles.
In third we have Bill de la T. Dauterive- who had a similar statline to Murphy, just a little less meat on the bone, still an amazing rookie campaign for the rookie LB.
Fourth on the list we have the Pizzaman himself- putting up solid weighted tackles and having decent alternate stats as well. Its really tough to put into numbers the value of a DT unless you are looking at a wild stat like run or pass block win rate but ultimately Pizzaman had a solid rookie season, sadly wasn't making the big plays that Reed or Murphy were.
Rounding out the top five is Fuzzy George who from the safety position really made an impact disruption the QB in the backfield with five sacks and like Reed was just making plays on the ball.
Ok now lets look at the single stat leaders and what they show…
TacklesPR
Murphy leads us here- Murphy ranked in the top third (14th/46) in LBs which is a big feat considering the talent that occupies this position on the big stage. Pizzaman was good for 4th/22 but with such a small field and slimmer top performers he takes a close second.
TFLPR
Taking the top slot by a pretty wide margin we have Owen Reed, and the big thing here is that he had half the TFLs for his position. You could be thinking, this is a little outlier based but thats still darn impressive. Pizzaman playing second fiddle where he ranked T5/22 impressive for a rook but he should be in the backfield- thats why he gets paid, while Reed is getting extra credit on these blitz schemes.
FFPR
Reed once again on top, this time closely followed by Will Smyth who's 2 FF actually kind of surprised me being ranked so high, but when you look at the breakdown, DTs were just not stripping the ball too much this season so him grabbing two was actually pretty ++.
SacksPR
Reed coming on top again here, alongside another CB Calvin Hobbes, once again this is mainly due to the position modifier for CBs, Reed lead the league for CBs with 5 and Hobbes was top 5. Hellraiser tops all rookies with 7 but as a DE he should be getting those numbers- finishing middle of the pack across all DEs.
INTPP
The total pizza points leader is back again here tied with a few other LBs for the top ranking. Small sample size here, but as we can see not many points awarded due to volume adjustments.
PDPR
The LBs taking the cake here again. CBs just get way too many opportunities to break up passes and Murphy having 8 is a real testament to his ability to smoke out the QBs targets in the middle of the field.
I'm very satisfied with these results overall for all the stats, certainly not perfect, but I think it does a great job of putting into perspective how different stats need to be weighed positionally. This took a little longer than I anticipated, but I'm glad I found a solid set of formulas to display my results in a thoughtful way rather than just plugging in numbers. I you made it this far, thank you and I hope you found this useful!
The main gist is instead of looking at raw stat totals, and weighing in my head how impactful this is compared across the board defensively- I created some formulas to spit out the proper leaders with weights baked in. I am taking the % of each stat compared to the total player pool with added weights based upon the frequency that each stat is attributed to a particular position. For example, LB tackles are weighted ~70% while DE tackles are ~190%, and CB TFL are weighted heavily almost 2000%. In order to account for randomness there is a weight based upon totals of the stat recorded. And just to add a cherry on top I added a stat modifier taking into account a simple impact of a stat say compared to a tackle (TFL= x2, FF = x3, Sack= x2, Int =x3, PD = x2)
I started this a week ago and kept just playing with some of the weights/ modifiers to a point where it kind of made sense. So here we are, a look at the top defensive rookies this season and a bit of analysis of the numbers. I’ll start with the overall leaders then we will breakdown each stat category leaders to look at what the system did good and what its weaknesses are. One small note is that I disregarded safeties (since no rookies recorded one) and did not take into account fumble recoveries as they are slightly random.
Here are the formulas that were used to compile these ratings, and the final pizza rating is just the sum of all the individual ratings.
PR=Raw Stat*Volume Modifier*Volume Weight*Stat Weight
where:
Raw stat= normal stat value
Volume modifier=# of players in position ÷ (position stat total*sqrt(positional adjustment))
Positional adjustment= avg stat per player ÷ (position stat total/# of players in position)
Volume weight=total stat #/10
stat weight= (T-x1, TFL= x2, FF = x3, Sack= x2, Int =x3, PD = x2)
Topping off the list we have Teddy Murphy who I think had probably the best all around season of the bunch- he was making plays, disrupting the QB and just an all around stud, a pretty damn good pizza rating of 205 with stellar ratings in tackles, INTs and PDs.
In second we have Murphy's teammate Owen Reed who is actually my front runner for DRoty, he had so much production from the CB position- solid ratings in tackles, amazing TFL representation and a standout with forced fumbles.
In third we have Bill de la T. Dauterive- who had a similar statline to Murphy, just a little less meat on the bone, still an amazing rookie campaign for the rookie LB.
Fourth on the list we have the Pizzaman himself- putting up solid weighted tackles and having decent alternate stats as well. Its really tough to put into numbers the value of a DT unless you are looking at a wild stat like run or pass block win rate but ultimately Pizzaman had a solid rookie season, sadly wasn't making the big plays that Reed or Murphy were.
Rounding out the top five is Fuzzy George who from the safety position really made an impact disruption the QB in the backfield with five sacks and like Reed was just making plays on the ball.
Ok now lets look at the single stat leaders and what they show…
TacklesPR
Murphy leads us here- Murphy ranked in the top third (14th/46) in LBs which is a big feat considering the talent that occupies this position on the big stage. Pizzaman was good for 4th/22 but with such a small field and slimmer top performers he takes a close second.
TFLPR
Taking the top slot by a pretty wide margin we have Owen Reed, and the big thing here is that he had half the TFLs for his position. You could be thinking, this is a little outlier based but thats still darn impressive. Pizzaman playing second fiddle where he ranked T5/22 impressive for a rook but he should be in the backfield- thats why he gets paid, while Reed is getting extra credit on these blitz schemes.
FFPR
Reed once again on top, this time closely followed by Will Smyth who's 2 FF actually kind of surprised me being ranked so high, but when you look at the breakdown, DTs were just not stripping the ball too much this season so him grabbing two was actually pretty ++.
SacksPR
Reed coming on top again here, alongside another CB Calvin Hobbes, once again this is mainly due to the position modifier for CBs, Reed lead the league for CBs with 5 and Hobbes was top 5. Hellraiser tops all rookies with 7 but as a DE he should be getting those numbers- finishing middle of the pack across all DEs.
INTPP
The total pizza points leader is back again here tied with a few other LBs for the top ranking. Small sample size here, but as we can see not many points awarded due to volume adjustments.
PDPR
The LBs taking the cake here again. CBs just get way too many opportunities to break up passes and Murphy having 8 is a real testament to his ability to smoke out the QBs targets in the middle of the field.
I'm very satisfied with these results overall for all the stats, certainly not perfect, but I think it does a great job of putting into perspective how different stats need to be weighed positionally. This took a little longer than I anticipated, but I'm glad I found a solid set of formulas to display my results in a thoughtful way rather than just plugging in numbers. I you made it this far, thank you and I hope you found this useful!
Code:
1044 words
The Riddler
New York Sliverbacks
Dallas Birddongs
Former Players
Dominos Pizzaman S30-S40
-------------------------------------------------------------------------------