|
Post by Chris Clement on Dec 31, 2015 19:34:30 GMT -6
Because those are {censored} things to waste your time on. If you know what you're doing you can derive real, tangible benefit, but it's a set of skills that's fairly rare. I don't mean to sound like quite as much of a braggart as this comes across, but what I do have are a very particular set of skills, skills I have acquired over a very long career. If you don't have those skills then don't worry about it, you'll waste a lot of time on the wrong stuff and get nothing for it.
|
|
|
Post by fantom on Dec 31, 2015 20:19:19 GMT -6
So all the guys on the nay side of this discussion are going to save time and energy by NOT doing stats in the coming season, right? Well, I don't do the stats so I don't care if we do them or not. I'm not really a naysayer on metrics if there's a metric that changes what we do. I don't agree with the "never punt" guy but he had stats to back up his theory and, even though most of us haven't bought in fully, I do think it's affected our thinking about punting near midfield in positive territory. I just don't think that the stuff that most people are talking about, stuff like goals charts, are a useful way to spend valuable time in-season.
|
|
|
Post by indian1 on Dec 31, 2015 21:12:05 GMT -6
I don't see a lot of value in scouring stats that tell me basically what I thought in the first place from watching the game film. I think the important issue is usually execution of things we already know we need to do.
Let's say we have a lot better chance to get off the field on D if we can hold a team to less than 4 yds per carry...so how do we do that? fit up better... shed blocks better...tackle better... All things we need to work on and be better at regardless of what that metric says.
|
|
|
Post by Chris Clement on Dec 31, 2015 21:48:26 GMT -6
Ok I can't really explain the difference between actionable and inactionable stats any better, I guess all I can say is that your brain sucks. Everyone's brain sucks. You're wrong about things ALL THE TIME. You see stuff, you think you know it, you're wrong. You introduce biases of all kinds, all the times. You like some players better than others, some players get the benefit of the doubt when you grade, or when you evaluate. Opponents have reputations that may or may not be accurate, and these things come into your thought process. You may be in a good or bad mood when you're working. You let high leverage plays influence your perception of certain calls. You fall victim to the hot hand fallacy. You may have fundamental misconceptions about how the game works from a God's eye view.
|
|
|
Post by 33coach on Dec 31, 2015 22:01:47 GMT -6
Ok I can't really explain the difference between actionable and inactionable stats any better, I guess all I can say is that your brain sucks. Everyone's brain sucks. You're wrong about things ALL THE TIME. You see stuff, you think you know it, you're wrong. You introduce biases of all kinds, all the times. You like some players better than others, some players get the benefit of the doubt when you grade, or when you evaluate. Opponents have reputations that may or may not be accurate, and these things come into your thought process. You may be in a good or bad mood when you're working. You let high leverage plays influence your perception of certain calls. You fall victim to the hot hand fallacy. You may have fundamental misconceptions about how the game works from a God's eye view. thats exactly it. have proof to back up your opinions. thats the power of analysis. where your eyes may fail, numbers dont.
|
|
|
Post by fantom on Dec 31, 2015 22:07:41 GMT -6
Ok I can't really explain the difference between actionable and inactionable stats any better, I guess all I can say is that your brain sucks. Everyone's brain sucks. You're wrong about things ALL THE TIME. You see stuff, you think you know it, you're wrong. You introduce biases of all kinds, all the times. You like some players better than others, some players get the benefit of the doubt when you grade, or when you evaluate. Opponents have reputations that may or may not be accurate, and these things come into your thought process. You may be in a good or bad mood when you're working. You let high leverage plays influence your perception of certain calls. You fall victim to the hot hand fallacy. You may have fundamental misconceptions about how the game works from a God's eye view. 'you?"
|
|
|
Post by Chris Clement on Dec 31, 2015 22:22:22 GMT -6
Oh, my brain sucks too.
|
|
|
Post by coachd5085 on Jan 1, 2016 11:11:45 GMT -6
Ok I can't really explain the difference between actionable and inactionable stats any better, I guess all I can say is that your brain sucks. Everyone's brain sucks. You're wrong about things ALL THE TIME. You see stuff, you think you know it, you're wrong. You introduce biases of all kinds, all the times. You like some players better than others, some players get the benefit of the doubt when you grade, or when you evaluate. Opponents have reputations that may or may not be accurate, and these things come into your thought process. You may be in a good or bad mood when you're working. You let high leverage plays influence your perception of certain calls. You fall victim to the hot hand fallacy. You may have fundamental misconceptions about how the game works from a God's eye view. Get that...my counter to it would be that once you obtain a large enough data set to reveal anything so deeply nuanced as to not be blatantly obvious nor "smoothed" over time (and therefore not influenced by outliers) .. wouldn't the data be unreliable as it wouldn't apply to those who would be creating the next data set? (IE--great data from last season doesn't apply to our new set of players this year playing against a new set of players on our opponents) For example, in the anecdote about the pass concept that didn't time up...didn't your OC/QB coach say "hey..why are we hitching here". Didn't the all 24 (all 22 for those of us in the states) show "hmm.. he needs to throw it NOW but its not there?"
|
|
|
Post by Chris Clement on Jan 2, 2016 12:27:00 GMT -6
To your first point, people grossly overestimate (I was once among them) the difference between teams. 8-0 teams aren't drastically better than 0-8 teams, they're just moderately better at everything they do. Any team sufficiently good or bad to invalidate the numbers will be so far off the charts that it won't matter.
To the second point, we probably should have, but it was a forest/trees thing. You see the idiosyncrasies of a play on the micro scale but you don't connect it to the macro. We knew the play was kind of a kludge, it was pre-R4 and didn't really look pretty, but we'd falsely associated it with success because we remembered Alex Vitt making great one-handers and Andrew Smith taking that hit but hanging on to put us in field goal range. We didn't remember all the incompletions because the slot couldn't get to the window on time. Those are mundane incompletions, they don't stick properly in your mind.
Bottom line, we were wrong. We screwed up. A classic flaw in human reasoning bit us in the rear and if we didn't write stuff down we wouldn't have noticed. It's not like I used advanced math to create some fancy formula, it was just division.
|
|
|
Post by spos21ram on Jan 2, 2016 13:04:44 GMT -6
To your first point, people grossly overestimate (I was once among them) the difference between teams. 8-0 teams aren't drastically better than 0-8 teams, they're just moderately better at everything they do. Any team sufficiently good or bad to invalidate the numbers will be so far off the charts that it won't matter. To the second point, we probably should have, but it was a forest/trees thing. You see the idiosyncrasies of a play on the micro scale but you don't connect it to the macro. We knew the play was kind of a kludge, it was pre-R4 and didn't really look pretty, but we'd falsely associated it with success because we remembered Alex Vitt making great one-handers and Andrew Smith taking that hit but hanging on to put us in field goal range. We didn't remember all the incompletions because the slot couldn't get to the window on time. Those are mundane incompletions, they don't stick properly in your mind. Bottom line, we were wrong. We screwed up. A classic flaw in human reasoning bit us in the rear and if we didn't write stuff down we wouldn't have noticed. It's not like I used advanced math to create some fancy formula, it was just division. Just to your first paragraph...what level are you talking about? I'd agree with you if you're talking NFL, but at the HS level there is a HUGE difference between an undefeated league champion and the last place team. Most 0-8 teams on their best day couldn't beat the best team playing their worst. Even in college....the difference between Alabama and UCF is astronomical. Sent from my SAMSUNG-SM-G900A using proboards
|
|
|
Post by Chris Clement on Jan 2, 2016 13:31:43 GMT -6
It's CIS football. The good teams are like good but not great FCS teams and the bad teams are like bad but not terrible FCS teams.
The 8-0 teams still beat the 0-8 teams by enormous margins, but the yards per play, the completion percentage, things like that, they're not astonishingly different. It's not like one team is 85% completion and one is 15%, it's the accretion of small advantages over a long game.
For something like first down probability (which I'll concede does involve some advanced use of excel if you don't want to pull your hair out doing it by hand, but the math is very very simple) it doesn't even matter if a team is an outlier, all that matters is the relative difference of two points. 1-20 is slightly better than 2-10, whether it's 67% vs 65% for an average team or 50% vs 47% for a bad team doesn't really matter when deciding whether or not to take the penalty.
Whether a P(1D) graph can be made for high school, I don't know, but if someone has a hudl with 400 or so games with the prescout done I'll give it a try and let you know. It would probably be a good bit noisier but I bet it would be tighter than you'd think. People use different schemes but the objective is still to make first downs, and there will be an inherent pressure toward doing so efficiently, even unintentionally, because it leads to scoring more points.
|
|
|
Post by jlenwood on Jan 3, 2016 11:35:27 GMT -6
Man.....this went on a lot longer than I thought it would.
|
|
red
Freshmen Member
Posts: 71
|
Post by red on Jan 6, 2016 13:43:21 GMT -6
We take ourselves too seriously. Figure out a way to measure the important things, the immeasurables . Are you and your staff good role models, are you a good mentor, do you grow your young coaches, do you use your program as a platform to do good in your community, are your players good citizens,....etc. Spend your off-season building relationships and being with your family. We all use stats, interpretations of those vary. Let me summarize all the metrics, when you have division 1 players all your numbers are better and you are a better coach. When you dont, your not. Culture is more important than scheme. Figure out a way to measure that.
|
|
|
Post by bignose on Jan 6, 2016 15:58:42 GMT -6
Did anybody notice that the Cleveland Browns just hired Paul DePodesta, an analytics maven, as "Chief Strategy officer?"
The power of data in making better decisions is his specialty. Not just X's and O's, but organizational decisions……...
|
|
|
Post by fantom on Jan 6, 2016 16:46:02 GMT -6
Did anybody notice that the Cleveland Browns just hired Paul DePodesta, an analytics maven, as "Chief Strategy officer?" The power of data in making better decisions is his specialty. Not just X's and O's, but organizational decisions……... Sure, The Browns are the model for doing things right.
|
|
|
Post by coachbdud on Jan 6, 2016 17:14:11 GMT -6
he's only there to advise Johnny Football when to hit or stay while he's playing blackjack in vegas
|
|
|
Post by coachphillip on Jan 7, 2016 9:15:07 GMT -6
I wonder what his hand efficiency is based on hits per table.
|
|
|
Post by fantom on Jan 7, 2016 20:32:43 GMT -6
Brian Billick was on one of the talk shows today. When he was asked about analytics he had two questions that I found interesting:
1. How can you make this information useful to me?
2. That math professor whop figured out that, statistically you should always go for it on 4th down, is he going to be standing on the sideline with me when we don't make it on our own 10?
|
|
|
Post by Chris Clement on Jan 7, 2016 21:31:22 GMT -6
So he's willing to lose a game in reality to avoid the perception of losing a game? And no, there's no math professor saying he should go for it every time. Professor Pulaski is welcome to publish whenever he wants, he'll deliver the keynote at Sloan if he does.
|
|
|
Post by silkyice on Jan 7, 2016 22:11:16 GMT -6
So he's willing to lose a game in reality to avoid the perception of losing a game? And no, there's no math professor saying he should go for it every time. Professor Pulaski is welcome to publish whenever he wants, he'll deliver the keynote at Sloan if he does. The problem here is what actually happens. Let's just say that the professor says you should go for it on 4th and 1 from your own 10. Well that really only works out correctly with 100's of attempts. You get those multiple attempts and multiple games in baseball and basketball. But you don't in football. So when you go for it and fail and then lose, then yes, the HEAD COACH will be held accountable. Everything Chip Kelly did might actually have been right. But his a$$ is fired.
|
|
|
Post by Chris Clement on Jan 7, 2016 22:12:59 GMT -6
Post hoc ergo propter hoc
|
|
jmg999
Junior Member
Posts: 263
|
Post by jmg999 on Jan 7, 2016 22:15:34 GMT -6
The paper that Billick was referring to was written by Dr. David Romer. In fact, Dr. Romer has actually published two papers, to my knowledge, that deal w/ the subject of kicking the ball on fourth down vs. going for it. What Billick did was the same thing the coach at Pulaski Academy did: They misinterpreted the results. Neither paper said to always go for it on fourth down. In actuality, there were very specific instances, where the expected value derived from going for it outweighed the expected value of kicking the ball. And, this was less science and more math, so he was able to mathematically prove that, under the parameters employed, there was statistical significance in not kicking in certain scenarios.
|
|
|
Post by coachd5085 on Jan 7, 2016 23:06:44 GMT -6
Post hoc ergo propter hoc I don't believe it is. Here is the deal: All of these type of "mathematically proven" examples of expected value or probability aren't as useful in things such as football as they are in cardgames because the actions creating the data set are not fixed by rather actions by DIFFERENT HUMAN BEINGS. That data set has been compiled from the results of interrelated actions of 22 individual match ups IN THE PAST and are not necessarily reflective of the match ups for the upcoming week or upcoming play. That said, I do appreciate the idea that such information at least creates a thought process for coaches, as opposed to just automatically doing things because thats "how they have been done"
|
|
|
Post by Chris Clement on Jan 7, 2016 23:18:17 GMT -6
That's just not true. I...can't put it any nicer than that. Football is not a special snowflake immune from math. Your team is not a special snowflake that deviates significantly from every other team out there. Past is prologue. This is hardly the only thing in the world that involves people and statistics.
Don't cross a multi lane street because one car waved you on, you're statistically very likely to get hit by a car on the other lane. Sure, people might act differently in this one scenario, but that doesn't make it a good idea.
The only defensible point Billick made was that he wants to deflect blame to avoid being fired. People's brains are so bad at math they'll blame the wrong guy!
Look, hockey coaches are also stupid. They wait until the game is basically over to pull the goalie. They need to pull him two or three times earlier to give themselves the best chance, but they don't. They don't because losing by two or three "looks worse." It's NOT worse. It's the same. You play to WIN the game, not to lose by a socially acceptable margin.
|
|
|
Post by fantom on Jan 7, 2016 23:20:08 GMT -6
Post hoc ergo propter hoc I don't believe it is. Here is the deal: All of these type of "mathematically proven" examples of expected value or probability aren't as useful in things such as football as they are in cardgames because the actions creating the data set are not fixed by rather actions by DIFFERENT HUMAN BEINGS. That data set has been compiled from the results of interrelated actions of 22 individual match ups IN THE PAST and are not necessarily reflective of the match ups for the upcoming week or upcoming play. That said, I do appreciate the idea that such information at least creates a thought process for coaches, as opposed to just automatically doing things because thats "how they have been done" As I've mentioned in past posts, this, used reasonably has changed what coaches actually do. It's noticeable how many times teams go for it on 4th down near midfield.
|
|
|
Post by fantom on Jan 7, 2016 23:22:01 GMT -6
That's just not true. I...can't put it any nicer than that. Football is not a special snowflake immune from math. Your team is not a special snowflake that deviates significantly from every other team out there. Past is prologue. This is hardly the only thing in the world that involves people and statistics. Don't cross a multi lane street because one car waved you on, you're statistically very likely to get hit by a car on the other lane. Sure, people might act differently in this one scenario, but that doesn't make it a good idea. The only defensible point Billick made was that he wants to deflect blame to avoid being fired. People's brains are so bad at math they'll blame the wrong guy! Look, hockey coaches are also stupid. They wait until the game is basically over to pull the goalie. They need to pull him two or three times earlier to give themselves the best chance, but they don't. They don't because losing by two or three "looks worse." It's NOT worse. It's the same. You play to WIN the game, not to lose by a socially acceptable margin. That might be true if coaches didn't get fired because of public perceptions.
|
|
|
Post by coachd5085 on Jan 7, 2016 23:36:33 GMT -6
That's just not true. I...can't put it any nicer than that. Football is not a special snowflake immune from math. Silky has posted several times how his "explosive play" data changed quite a bit the years prior to his running back who signed with (I believe) Nebraska, the years with his RB who signed with Nebraska, and the years after the departure. Just one quick example off the top of my head that has already been given. MATH...says the average overnight low in January for New Orleans should be about 44 degrees. It will be around 60 degrees tomorrow night. The MATH that you are claiming is "not immunue" gives you data that came from plays where my RG blocked against previous opponents, not the upcoming opponent.
|
|
jmg999
Junior Member
Posts: 263
|
Post by jmg999 on Jan 7, 2016 23:40:04 GMT -6
That's just not true. I...can't put it any nicer than that. Football is not a special snowflake immune from math. Your team is not a special snowflake that deviates significantly from every other team out there. Past is prologue. This is hardly the only thing in the world that involves people and statistics. Don't cross a multi lane street because one car waved you on, you're statistically very likely to get hit by a car on the other lane. Sure, people might act differently in this one scenario, but that doesn't make it a good idea. The only defensible point Billick made was that he wants to deflect blame to avoid being fired. People's brains are so bad at math they'll blame the wrong guy! Look, hockey coaches are also stupid. They wait until the game is basically over to pull the goalie. They need to pull him two or three times earlier to give themselves the best chance, but they don't. They don't because losing by two or three "looks worse." It's NOT worse. It's the same. You play to WIN the game, not to lose by a socially acceptable margin. That might be true if coaches didn't get fired because of public perceptions. The same applies in any field. CEOs of major corporations get fired b/c of public perception. That doesn't negate the fact that they utilize statistical models to help manage their business platforms.
|
|
jmg999
Junior Member
Posts: 263
|
Post by jmg999 on Jan 7, 2016 23:42:05 GMT -6
That's just not true. I...can't put it any nicer than that. Football is not a special snowflake immune from math. Silky has posted several times how his "explosive play" data changed quite a bit the years prior to his running back who signed with (I believe) Nebraska, the years with his RB who signed with Nebraska, and the years after the departure. Just one quick example off the top of my head that has already been given. MATH...says the average overnight low in January for New Orleans should be about 44 degrees. It will be around 60 degrees tomorrow night. The MATH that you are claiming is "not immunue" gives you data that came from plays where my RG blocked against previous opponents, not the upcoming opponent. You seem to be under the assumption that these statistical models are somehow predicting the future. They're not. They're not even trying to do so. They're simply using models in order to help sort through the randomness.
|
|
|
Post by silkyice on Jan 8, 2016 8:16:34 GMT -6
The data might show that on 4th and 1 inch, sneak is by far the best play. Sounds great. But what happens when my backup 165 pound center is in and playing against a 300 pound noseguard who has been offered by Bama?
There ain't no data for that.
|
|