TheElectricCompany Posted October 6, 2020 Posted October 6, 2020 (edited) 7 minutes ago, MJS said: All the other fake INT's? Pretty sure that number is zero. I thought it was a bad call too, but its a INT. My point is that we're going down the asterisk path, then those conditions need to be applied across the entire database. Edited October 6, 2020 by TheElectricCompany
TBBills Posted October 6, 2020 Posted October 6, 2020 Just now, TheElectricCompany said: I thought it was a bad call too, but its a INT. My point is that we're going down the asterisk path, then those conditions need to be applied across the entire database. They are which is why Josh Allen is the second best in the league right now.
MJS Posted October 6, 2020 Posted October 6, 2020 1 minute ago, TheElectricCompany said: I thought it was a bad call too, but its a INT. My point is that we're going down the asterisk path, then those conditions need to be applied across the entire database. I was simply curious what the score would be.
BuffaloHokie13 Posted October 6, 2020 Author Posted October 6, 2020 6 minutes ago, Mr. WEO said: Well by that reasoning, every pass is a "loss of control of the ball by the QB".... But no, it's just a bad pass. QB has the ability (by definition) to control where he throws the ball. He has little or no control as to where the ball goes after he fumbles. Let's put this another way: One QB fumbles 50 times and his team recovers 48 times so he has only "2 turnovers" Another QB fumbles 50 times and his team can only recover 5, so he has "45 turnovers" Which QB is better at protecting the ball? 2
whatdrought Posted October 6, 2020 Posted October 6, 2020 14 minutes ago, Mr. WEO said: Well by that reasoning, every pass is a "loss of control of the ball by the QB".... But no, it's just a bad pass. QB has the ability (by definition) to control where he throws the ball. He has little or no control as to where the ball goes after he fumbles. Right, and both a bad pass and a fumble are bad things - no question about it. The difference is when looking at the outcome. Statistics are based on outcomes. When a WR drops a pass you say "Josh completed 20 0f 25 passes for 301 yards, but the recievers dropped the ball 3 times." The implication is that the outcome could have been different (in this case, better) but we can't determine what it was. We don't say "because of where those drops occured, we can add 75 yards to his passing total." It's all about outcome. So we can say "He fumbled twice, and lost one for one turnover." Which places him on the same level as the QB who threw one interception. The outcome (the turnover) is the important part.
GoBills808 Posted October 6, 2020 Posted October 6, 2020 1 hour ago, Mr. WEO said: Who recovers a fumbled ball is often random--whoever happens to be nearest it during this unexpected event. If a QB fumbles a lot--that's a problem. Can't really argue that, because he is lucky enough to have the fumbled ball bounce his team's way, that it's "OK". Bad passes are either caught or dropped, and recorded as such. There's no "potential fumble", so there is no need to score "potential ints". Further, I believe there's evidence to suggest that INTs, while still beholden to an element of luck, can be influenced by a measure of skill on the defender's side, whereas fumble recoveries by one team or another are almost completely if not entirely coincidental. 1
GoBills808 Posted October 6, 2020 Posted October 6, 2020 10 minutes ago, whatdrought said: Right, and both a bad pass and a fumble are bad things - no question about it. The difference is when looking at the outcome. Statistics are based on outcomes. When a WR drops a pass you say "Josh completed 20 0f 25 passes for 301 yards, but the recievers dropped the ball 3 times." The implication is that the outcome could have been different (in this case, better) but we can't determine what it was. We don't say "because of where those drops occured, we can add 75 yards to his passing total." It's all about outcome. So we can say "He fumbled twice, and lost one for one turnover." Which places him on the same level as the QB who threw one interception. The outcome (the turnover) is the important part. This is incorrect. In poker we'd say you are being results-oriented instead of playing how the math dictates. 1
Mr. WEO Posted October 6, 2020 Posted October 6, 2020 (edited) 22 minutes ago, whatdrought said: Right, and both a bad pass and a fumble are bad things - no question about it. The difference is when looking at the outcome. Statistics are based on outcomes. When a WR drops a pass you say "Josh completed 20 0f 25 passes for 301 yards, but the recievers dropped the ball 3 times." The implication is that the outcome could have been different (in this case, better) but we can't determine what it was. We don't say "because of where those drops occured, we can add 75 yards to his passing total." It's all about outcome. So we can say "He fumbled twice, and lost one for one turnover." Which places him on the same level as the QB who threw one interception. The outcome (the turnover) is the important part. Dropped passes are assigned to the receivers. Dropped INTs are not assigned to a Defender. Both are simply incomplete passes assigned to the QB. The QB who fumbles a lot is poor at ball protection, just as is a QB who tosses a lot of ints. Difference is that in a fumble, the QB does not intend to lose control of the ball whereas in a pass, by definition, that's exactly what he does (as you said upstream). A completed pass is not recorded as "a missed int". Once the QB fumbles, it's pure luck as to who it bounces to. A caught pass and an INT are not typically consider the outcomes of luck or chance. The QB, the receiver and even the Defender have significant control of what happens after the ball leaves the hand of the QB on a pass. None of the is true with a QB fumble. Edited October 6, 2020 by Mr. WEO
whatdrought Posted October 6, 2020 Posted October 6, 2020 (edited) 27 minutes ago, Mr. WEO said: Dropped passes are assigned to the receivers. Dropped INTs are not assigned to a Defender. Both are simply incomplete passes assigned to the QB. The QB who fumbles a lot is poor at ball protection, just as is a QB who tosses a lot of ints. Difference is that in a fumble, the QB does not intend to lose control of the ball whereas in a pass, by definition, that's exactly what he does (as you said upstream). A completed pass is not recorded as "a missed int". Once the QB fumbles, it's pure luck as to who it bounces to. A caught pass and an INT are not typically consider the outcomes of luck or chance. The QB, the receiver and even the Defender have significant control of what happens after the ball leaves the hand of the QB on a pass. None of the is true with a QB fumble. But there does exist luck when a QB throws a ball in the direction of the DB who drops it... That's what I'm saying. I'm not excusing non-lost fumbles or saying they aren't a problem. I'm saying that to truly track turnovers you have to look at one key thing -- did the ball get turned over. A fumble is a potential turnover and you're absolutely right, it's a negative stat towards the QB and it's all a matter of luck from then forward as to whether or not it becomes a TO. But we don't say "QB A has 20 turnovers compared to QB B's 6" when 15 of those are fumbles that the offense recovered - just like we wouldn't say that about passes that bounce off the DB's chest. From the outcome perspective the luck has been factored in and we count TO's as the true comparison grid. 34 minutes ago, GoBills808 said: This is incorrect. In poker we'd say you are being results-oriented instead of playing how the math dictates. Could you further explain this? Not sure I get what you're saying. Edited October 6, 2020 by whatdrought
GoBills808 Posted October 6, 2020 Posted October 6, 2020 7 minutes ago, whatdrought said: Could you further explain this? Not sure I get what you're saying. EV (expected value) in cards is similar to EPA wherein each action or play is either positive or negative (or neutral) depending on a set of known variables. This is quantifiable math as opposed to DVOA/PFF grades which are proprietary and/or subjective. Example- say the EV of a $50 bet on a particular board, with a particular hand against a range of particular hands your opponent could have, with respect to your implied odds of winning X dollars is POSITIVE, or +EV...if you are playing a mathematically sound game you should always make that bet regardless of outcome. You'll lose the hand a nonzero percentage but if a play is +EV it is the correct move, no matter if you end up winning or losing. Similar to QB fumble: a QB fumble is always -EPA regardless of who happens to recover it. A bad pass thrown at the defender’s chest is -EPA regardless of whether he intercepts it or it falls to the ground. Using outcomes to determine whether a particular action is positive or negative is inherently flawed analysis, statistically speaking.
Playoffs? Posted October 6, 2020 Posted October 6, 2020 Love seeing Baker Mayfield so low. Just not a fan of his. And his crop of WR’s are top of the line. Wonder how low his rating has to be before they cut a few of his commercials!
Mr. WEO Posted October 6, 2020 Posted October 6, 2020 1 hour ago, whatdrought said: But there does exist luck when a QB throws a ball in the direction of the DB who drops it... That's what I'm saying. I'm not excusing non-lost fumbles or saying they aren't a problem. I'm saying that to truly track turnovers you have to look at one key thing -- did the ball get turned over. A fumble is a potential turnover and you're absolutely right, it's a negative stat towards the QB and it's all a matter of luck from then forward as to whether or not it becomes a TO. But we don't say "QB A has 20 turnovers compared to QB B's 6" when 15 of those are fumbles that the offense recovered - just like we wouldn't say that about passes that bounce off the DB's chest. From the outcome perspective the luck has been factored in and we count TO's as the true comparison grid. Could you further explain this? Not sure I get what you're saying. Again, by your logic every pass is a potential into and turnover. The problem with that is obvious. The QB is purposely giving up the ball as a pass. If he wants to avoid a sack or bad throw he can toss it out of bounds and prevent a bad outcome. Not so for a fumble. A fumble is an inability to safely possess the ball by the QB. Therefore it is a legitimate stat. A QB who fumbles a lot is a liability. One who fumbles a lot and he or is team happens to recover a lot of them is lucky, not less of a liability
No_Matter_What Posted October 6, 2020 Posted October 6, 2020 OP I really like your rankings. Seems most accurate from any rankings based on stats. I don't watch enough games to have opinion on everybody and it is hard to rank QBs, but your list meets my eyetest. Could you please post your final rankings from last year (or link the thread) here?
FireChans Posted October 7, 2020 Posted October 7, 2020 10 hours ago, BuffaloHokie13 said: As many of you know and some will learn now, I set out a few years back to develop a comparative rating system that, when sorted, would more accurately line up with the eye test than traditional stats. I enter all of the data manually each week, so if you see an error please let me know! Common questions: - fumbles refers specifically to fumbles lost, as it is factored into total turnovers - yards per touch (YPT) is the QB's pass and rush yards, minus their sack yards, divided by their total attempts - the final column is my current ratings - not ESPN's garbage QBR stat - click on the image to make it bigger. I know there's a lot of info there. - cutoff to qualify was a minimum of 75 pass attempts Anyway, let me know what you think of the order through the first quarter of the season and let's discuss! Go Bills! I remember when we would debate Tyrod at 19 or 22 only a few short years ago. 1
Ralonzo Posted October 7, 2020 Posted October 7, 2020 18 hours ago, whatdrought said: Could you further explain this? Not sure I get what you're saying. I’d guess it’s a reference to the fallacy of betting big when you have slim chance of winning a hand, but then your lucky card comes and you win the pot. You mistakenly concluded that betting was the proper play rather than folding, but over time the percentages are against you so you will lose more than win.
No_Matter_What Posted November 2, 2020 Posted November 2, 2020 Hi @BuffaloHokie13, waiting for and looking forward your Q2 rankings. I don't know how hard it is for you, but it would be great if you could make Q2 chart separately as well as first half totals.
BuffaloHokie13 Posted November 2, 2020 Author Posted November 2, 2020 15 hours ago, No_Matter_What said: Hi @BuffaloHokie13, waiting for and looking forward your Q2 rankings. I don't know how hard it is for you, but it would be great if you could make Q2 chart separately as well as first half totals. Based on byes and attempts I'm looking at week 9 for Q2. I'm not really set up for individual quarters, as I overwrite my week to week stats - no reason here other than tab management. But if I find some extra time on the weekend then that breakout shouldn't be too difficult. Cheers! 3
No_Matter_What Posted November 2, 2020 Posted November 2, 2020 35 minutes ago, BuffaloHokie13 said: Based on byes and attempts I'm looking at week 9 for Q2. I'm not really set up for individual quarters, as I overwrite my week to week stats - no reason here other than tab management. But if I find some extra time on the weekend then that breakout shouldn't be too difficult. Cheers! Thanks man. I actually thought you might be doing first half after W2. If you find time to separate Q2 great, if not thanks for effort anyway.
Recommended Posts