|
Post by Chris Clement on Feb 1, 2015 1:24:56 GMT -6
I have access to a fair bit of data from a pretty competitive conference and I've found many descriptive statistics, but predictive statistics are more challenging. I've built a pretty good P(1D) model, that shows probability of converting a given down and distance. It tells you when you should accept or decline penalties like 1-5 vs 2-short and creates a better definition of what constitutes a successful play.
|
|
|
Post by silkyice on Feb 1, 2015 7:42:30 GMT -6
I have access to a fair bit of data from a pretty competitive conference and I've found many descriptive statistics, but predictive statistics are more challenging. I've built a pretty good P(1D) model, that shows probability of converting a given down and distance. It tells you when you should accept or decline penalties like 1-5 vs 2-short and creates a better definition of what constitutes a successful play. Now that is something that can be used as a guide. Unlike, when we rush for 300 yards we win 90% of the time, therefore our goal is to rush for 300 yards. Well no crap, when you rush for 300 in high school, you usually are the better team. The point is having rushing 300 as a goal does nothing. Deciding to work your tail off in the weight room and coach better so that you can rush for 300 means something. But who isn't already trying to do that? So here is my question about goals and moneyball bull crap as it relates to high school football. You are down 6 points with 30 seconds left and it is 4th and 10 and the ball is on the 50 and you have 299 yards rushing. Should you try and throw to pick up the first down or run sneak to make sure you get your 300 yards and 90% winning percentage?
|
|
|
Post by blb on Feb 1, 2015 7:51:56 GMT -6
How many of us are actually aware of what our stats are during a game (other than the obvious turnovers)?
How would that knowledge help the kids to play better?
Even if you did know say, at halftime, wouldn't the "how" of achieving pre-established goals be more important than the numbers themselves?
|
|
|
Post by natenator on Feb 1, 2015 8:03:00 GMT -6
What was alluded to early is that moneyball was about finding value in players who were perceived as low value.
Most of us cannot do that. We don't really get to pick and choose. I coach a rep team (regional 'all-star' team for those not aware what rep is) and I still don't get to really pick and choose. I may have 60 players tryout but most of those 60 will make the team as long as they can stand up - how much they play is a different story.
But either way, moneyball was about selecting players who had statistics which correlated to certain factors that felt contributed to wins.
|
|
|
Post by silkyice on Feb 1, 2015 8:38:23 GMT -6
Another point about Moneyball and how it relates to football, is that the goal is almost always to get on base in baseball. No matter if you are up 1 or up 10, down 1 or down 10, no matter what inning it is, you help your team when up to bat by getting on base.
In football, the down, distance, field position, score, time remaining, defensive structure, personnell, etc., all play a role in what you should call or attempt to do.
|
|
|
Post by ryanculloty on Feb 1, 2015 8:43:45 GMT -6
Great stuff--thanks coach Really? Not to demean coachweav88 's link, but I wouldn't say anything mentioned in that link would be "great". I read it and I thought--hmm..so Tony Demeo suggests coaches be good coaches. I mean, some of the ideas presented such as : -"Be fundamentally sound, good tacklers and play fast" -"Gang Tackle and strip the ball on every tackle once the tackle is secure. Touch the ball on defense on every play." are fairly nice football concepts, but I wouldn't say they are geared to any type of metric or statistical analysis. It is just good football. I hope this doesn't come off like I am knocking Coach Demeo, because he is obviously a fine football coach. I just think we as coaches may do ourselves a disservice chasing after "moneyball" solutions, particularly ones tied to the metrics and stats Coach Demeo espouses (Score 25 a game against the best, and hold best opponents to 16) That said, I DO think that the most underlying concept of "moneyball" --THINK ABOUT YOUR SITUATION, AND COME UP WITH A PLAN-- can be useful. I believe this was discussed in another thread about always going for it, always onside kicking etc. Point of reference...I like things in the link...not everything....but enough to get the wheels turning... solid points here coach--thanks
|
|
|
Post by blb on Feb 1, 2015 8:54:47 GMT -6
George Carlin gave the best explanation of the differences between baseball and football (YouTube or Google it - classic).
Seriously, the biggest difference is football is timed, baseball is not.
|
|
|
Post by coachd5085 on Feb 1, 2015 11:06:01 GMT -6
George Carlin gave the best explanation of the differences between baseball and football (YouTube or Google it - classic). Seriously, the biggest difference is football is timed, baseball is not. I think in the case of trying to discover and apply evaluation metrics--the biggest difference is that football has a lot more interrelated events than baseball.
|
|
|
Post by 33coach on Feb 1, 2015 11:39:30 GMT -6
George Carlin gave the best explanation of the differences between baseball and football (YouTube or Google it - classic). Seriously, the biggest difference is football is timed, baseball is not. I think in the case of trying to discover and apply evaluation metrics--the biggest difference is that football has a lot more interrelated events than baseball. your right, the game is more complex. which is why i gave up pretty quickly on trying to variablize & quantify the game. math is useful after the fact; i can say that in the games that i have stats for, if our First Down Percentage per Drive was over 45% (5 first downs in 12 play drive) we had a 90% win rate. but that doesnt tell me how to get a First Down Percentage per Drive over 45% next week....
|
|
|
Post by rsmith627 on Feb 1, 2015 12:13:46 GMT -6
I think in the case of trying to discover and apply evaluation metrics--the biggest difference is that football has a lot more interrelated events than baseball. your right, the game is more complex. which is why i gave up pretty quickly on trying to variablize & quantify the game. math is useful after the fact; i can say that in the games that i have stats for, if our First Down Percentage per Drive was over 45% (5 first downs in 12 play drive) we had a 90% win rate. but that doesnt tell me how to get a First Down Percentage per Drive over 45% next week.... No, but you can figure out how you picked up those first downs, and which plays probably need to come out of your playbook forever. I coached at a place a few years back where we had a few pass concepts that never did anything for us, yet we kept calling them over and over and over again.
|
|
|
Post by fantom on Feb 1, 2015 12:21:43 GMT -6
your right, the game is more complex. which is why i gave up pretty quickly on trying to variablize & quantify the game. math is useful after the fact; i can say that in the games that i have stats for, if our First Down Percentage per Drive was over 45% (5 first downs in 12 play drive) we had a 90% win rate. but that doesnt tell me how to get a First Down Percentage per Drive over 45% next week.... No, but you can figure out how you picked up those first downs, and which plays probably need to come out of your playbook forever. I coached at a place a few years back where we had a few pass concepts that never did anything for us, yet we kept calling them over and over and over again. Forever is a long time. There are a lot of variables: Different players, different opponents, different situations. What doesn't work this week might work next week or next year. Or the other way around.
|
|
|
Post by coachd5085 on Feb 1, 2015 12:21:56 GMT -6
your right, the game is more complex. which is why i gave up pretty quickly on trying to variablize & quantify the game. math is useful after the fact; i can say that in the games that i have stats for, if our First Down Percentage per Drive was over 45% (5 first downs in 12 play drive) we had a 90% win rate. but that doesnt tell me how to get a First Down Percentage per Drive over 45% next week.... Exactly. Stats like that are just further indicators that you whipped up on the opponent. Restated--it says "90% of the time we kick the other teams behind...we win" As you state, those stats don't tell you how to obtain such statistics.
|
|
|
Post by rsmith627 on Feb 1, 2015 12:25:31 GMT -6
No, but you can figure out how you picked up those first downs, and which plays probably need to come out of your playbook forever. I coached at a place a few years back where we had a few pass concepts that never did anything for us, yet we kept calling them over and over and over again. Forever is a long time. There are a lot of variables: Different players, different opponents, different situations. What doesn't work this week might work next week or next year. Or the other way around. Forever is an exaggeration, but the point stands. We had a play action concept that was generally really good to us, but our opponent had it well scouted and blew it up every time we ran it. I want to say we called that play 9-10 times that night. I have no idea why we kept going to it. This happened a couple of weeks in a row. Our OC wasn't a stat guy and didn't look at data like this, which probably would have been useful.
|
|
|
Post by 33coach on Feb 1, 2015 12:26:16 GMT -6
your right, the game is more complex. which is why i gave up pretty quickly on trying to variablize & quantify the game. math is useful after the fact; i can say that in the games that i have stats for, if our First Down Percentage per Drive was over 45% (5 first downs in 12 play drive) we had a 90% win rate. but that doesnt tell me how to get a First Down Percentage per Drive over 45% next week.... No, but you can figure out how you picked up those first downs, and which plays probably need to come out of your playbook forever. I coached at a place a few years back where we had a few pass concepts that never did anything for us, yet we kept calling them over and over and over again. i can get behind part of that. but its still retrospective, the thing about sabermetrics is that its predictive...i just dont think there is anything like that for football...too many variables to keep track of. looking more and more into this, you gain a TON of respect for the developers at EA Sports creating simulation football games.
|
|
|
Post by fantom on Feb 1, 2015 12:28:42 GMT -6
Forever is a long time. There are a lot of variables: Different players, different opponents, different situations. What doesn't work this week might work next week or next year. Or the other way around. Forever is an exaggeration, but the point stands. We had a play action concept that was generally really good to us, but our opponent had it well scouted and blew it up every time we ran it. I want to say we called that play 9-10 times that night. I have no idea why we kept going to it. This happened a couple of weeks in a row. Our OC wasn't a stat guy and didn't look at data like this, which probably would have been useful. Data? After I call something twice and it doesn't work it's done for the day. I don't need to crunch the numbers there.
|
|
|
Post by rsmith627 on Feb 1, 2015 12:29:29 GMT -6
Forever is an exaggeration, but the point stands. We had a play action concept that was generally really good to us, but our opponent had it well scouted and blew it up every time we ran it. I want to say we called that play 9-10 times that night. I have no idea why we kept going to it. This happened a couple of weeks in a row. Our OC wasn't a stat guy and didn't look at data like this, which probably would have been useful. Data? After I call something twice and it doesn't work it's done for the day. I don't need to crunch the numbers there. LOL you would think...needless to say I have moved on from that program.
|
|
|
Post by silkyice on Feb 1, 2015 12:34:00 GMT -6
Forever is a long time. There are a lot of variables: Different players, different opponents, different situations. What doesn't work this week might work next week or next year. Or the other way around. Forever is an exaggeration, but the point stands. We had a play action concept that was generally really good to us, but our opponent had it well scouted and blew it up every time we ran it. I want to say we called that play 9-10 times that night. I have no idea why we kept going to it. This happened a couple of weeks in a row. Our OC wasn't a stat guy and didn't look at data like this, which probably would have been useful. Do you really need stats to know a play isn't working at all?
|
|
|
Post by rsmith627 on Feb 1, 2015 12:37:07 GMT -6
Forever is an exaggeration, but the point stands. We had a play action concept that was generally really good to us, but our opponent had it well scouted and blew it up every time we ran it. I want to say we called that play 9-10 times that night. I have no idea why we kept going to it. This happened a couple of weeks in a row. Our OC wasn't a stat guy and didn't look at data like this, which probably would have been useful. Do you really need stats to know a play isn't working at all? Do I? No. This guy needed something though, because common sense was lacking.
|
|
|
Post by ryanculloty on Feb 1, 2015 12:47:17 GMT -6
your right, the game is more complex. which is why i gave up pretty quickly on trying to variablize & quantify the game. math is useful after the fact; i can say that in the games that i have stats for, if our First Down Percentage per Drive was over 45% (5 first downs in 12 play drive) we had a 90% win rate. but that doesnt tell me how to get a First Down Percentage per Drive over 45% next week.... Exactly. Stats like that are just further indicators that you whipped up on the opponent. Restated--it says "90% of the time we kick the other teams behind...we win" As you state, those stats don't tell you how to obtain such statistics. Even if it is a "retro look" at things--seeing how effective you are running your plays can be extremely valuable Plays shut down due to 1. The opponent's play 2. not proper play technique #1--as established--is often a great variable and must be addressed individually each week #2--This item can help to mold a better and more effective practice plan the following week--thus helping you improve clean-up rule interpretation--spend more time on specific techniques to make the play go--look a personal--play rep v. various look all will help to make you better--thus a stat that I see as having true value The other "stat" an old coach told me--not sure how to chart it was-- Find out who your guys are that can put the ball in the end zone and be sure to dial up their number a lot... That might not go with this discussion but it has a lot of value.
|
|
|
Post by coachd5085 on Feb 1, 2015 12:54:11 GMT -6
Exactly. Stats like that are just further indicators that you whipped up on the opponent. Restated--it says "90% of the time we kick the other teams behind...we win" As you state, those stats don't tell you how to obtain such statistics. Even if it is a "retro look" at things--seeing how effective you are running your plays can be extremely valuable Plays shut down due to 1. The opponent's play 2. not proper play technique #1--as established--is often a great variable and must be addressed individually each week #2--This item can help to mold a better and more effective practice plan the following week--thus helping you improve clean-up rule interpretation--spend more time on specific techniques to make the play go--look a personal--play rep v. various look all will help to make you better--thus a stat that I see as having true value The other "stat" an old coach told me--not sure how to chart it was-- Find out who your guys are that can put the ball in the end zone and be sure to dial up their number a lot... That might not go with this discussion but it has a lot of value. Coach-- what you are describing would be observed via film study. No need to crunch numbers.
|
|
|
Post by ryanculloty on Feb 1, 2015 13:23:57 GMT -6
Even if it is a "retro look" at things--seeing how effective you are running your plays can be extremely valuable Plays shut down due to 1. The opponent's play 2. not proper play technique #1--as established--is often a great variable and must be addressed individually each week #2--This item can help to mold a better and more effective practice plan the following week--thus helping you improve clean-up rule interpretation--spend more time on specific techniques to make the play go--look a personal--play rep v. various look all will help to make you better--thus a stat that I see as having true value The other "stat" an old coach told me--not sure how to chart it was-- Find out who your guys are that can put the ball in the end zone and be sure to dial up their number a lot... That might not go with this discussion but it has a lot of value. Coach-- what you are describing would be observed via film study. No need to crunch numbers. I understand what you are stating but again it is the an opportunity to connect a staff-players--both with an achievement goal Let's say that as a staff you sit down and determine that for you to be successful you need to achieve a success rate of 75% on your plays... That now becomes the tracking stat--the kids buy into more as something they want to achieve & how it connects to winning games What then needs to be established is what = efficient? I would envision plays gaining 4+ yards on 1-2 down as efficient plays that get 1st downs as efficient and plays that scored as efficient Turn-overs Play efficiency These 2 look to me as a big factor The other thing to remember that yes, there are examples we all can point to where the W didn't come and we achieved "X" However, the focus to achieve X creates the process---and the process leads to consistency and continued growth hence value my 2 cents.
|
|
|
Post by fantom on Feb 1, 2015 14:30:41 GMT -6
Coach-- what you are describing would be observed via film study. No need to crunch numbers. I understand what you are stating but again it is the an opportunity to connect a staff-players--both with an achievement goal Let's say that as a staff you sit down and determine that for you to be successful you need to achieve a success rate of 75% on your plays... That now becomes the tracking stat--the kids buy into more as something they want to achieve & how it connects to winning games What then needs to be established is what = efficient? I would envision plays gaining 4+ yards on 1-2 down as efficient plays that get 1st downs as efficient and plays that scored as efficient Turn-overs Play efficiency These 2 look to me as a big factor The other thing to remember that yes, there are examples we all can point to where the W didn't come and we achieved "X" However, the focus to achieve X creates the process---and the process leads to consistency and continued growth hence value my 2 cents. Efficient against whom? If a play gains 1 YPC against the state champs and 10 the next week against J.C. Penney Mannequins it averages out as efficient.
|
|
|
Post by blb on Feb 1, 2015 14:47:36 GMT -6
When I started coaching the prevailing offensive wisdom was to control the ball as long as possible (TOP), attempt to gain field position that way, avoid negative plays including penalties and turnovers.
The defensive thinking was force the offense to drive the ball, go the "long hard way" and wait for them to make a mistake.
To me that was contradictory.
What I have learned is big plays on offense, defense, Special Teams win games, especially between evenly matched teams.
Typically there are 5-6 such plays a game.
|
|
|
Post by ryanculloty on Feb 1, 2015 14:52:34 GMT -6
I understand what you are stating but again it is the an opportunity to connect a staff-players--both with an achievement goal Let's say that as a staff you sit down and determine that for you to be successful you need to achieve a success rate of 75% on your plays... That now becomes the tracking stat--the kids buy into more as something they want to achieve & how it connects to winning games What then needs to be established is what = efficient? I would envision plays gaining 4+ yards on 1-2 down as efficient plays that get 1st downs as efficient and plays that scored as efficient Turn-overs Play efficiency These 2 look to me as a big factor The other thing to remember that yes, there are examples we all can point to where the W didn't come and we achieved "X" However, the focus to achieve X creates the process---and the process leads to consistency and continued growth hence value my 2 cents. Efficient against whom? If a play gains 1 YPC against the state champs and 10 the next week against J.C. Penney Mannequins it averages out as efficient. It would go to my above point---plays are not efficient due to 1. who you are playing 2. your own error You cannot always can control #1 or must game plan to be successful Where #2 can help you build a better practice plan and continue to perfect your plays
|
|
|
Post by fantom on Feb 1, 2015 15:02:37 GMT -6
Efficient against whom? If a play gains 1 YPC against the state champs and 10 the next week against J.C. Penney Mannequins it averages out as efficient. It would go to my above point---plays are not efficient due to 1. who you are playing 2. your own error You cannot always can control #1 or must game plan to be successful Where #2 can help you build a better practice plan and continue to perfect your plays If you think it helps you, and maybe it does, keep doing it.
|
|
|
Post by ryanculloty on Feb 1, 2015 15:11:45 GMT -6
It would go to my above point---plays are not efficient due to 1. who you are playing 2. your own error You cannot always can control #1 or must game plan to be successful Where #2 can help you build a better practice plan and continue to perfect your plays If you think it helps you, and maybe it does, keep doing it. Fantom What do you suggest? Appreciate the feedback--the discussion helps me-- I guess Football, like education, has a ton of factors that bring or prevent success-- I know when our district started looking a some pieces of data we improved dramatically--it brought focus and clarity of purpose for everyone---were the factors of focus the only things that brought success---no but it brought us together to work toward a common goal and the results were positive. Are there other factors--yes--but without asking the question we would have never improved in the ways we have as a top school in the county.... I am trying to find what works with this line of thinking for football...and people like you have tons of experience and knowledge to share and exchange I appreciate your insight and assistance!
|
|
|
Post by ryanculloty on Feb 1, 2015 15:14:35 GMT -6
When I started coaching the prevailing offensive wisdom was to control the ball as long as possible (TOP), attempt to gain field position that way, avoid negative plays including penalties and turnovers.
The defensive thinking was force the offense to drive the ball, go the "long hard way" and wait for them to make a mistake.
To me that was contradictory.
What I have learned is big plays on offense, defense, Special Teams win games, especially between evenly matched teams.
Typically there are 5-6 such plays a game. blb Great points--when I was a DC--I always charted the opponents Big Plays and did everything I could to prevent them--especially if those big plays paired up with what I thought were their top plays in general Key points--thank you!
|
|
|
Post by blb on Feb 1, 2015 15:41:44 GMT -6
TV analysts (including those we'll see tonight) make hundreds of thousands of dollars for saying every week, "The team that makes the fewest (or gets the most) turnovers will win."
And-or makes the most big plays.
Well, duh.
They have no answer on how to do that anymore than we do though, other than play fast, hard, smart, and efficiently.
Usually the team with the best players is able to do that the most.
|
|
|
Post by coachd5085 on Feb 1, 2015 15:44:55 GMT -6
Let's say that as a staff you sit down and determine that for you to be successful you need to achieve a success rate of 75% on your plays... Why? I want a success rate of 100% Obviously that won't happen. But what do I gain from sitting down and setting some other arbitrary success rate? The problem with trying to install a TQM or Six Sigma styled metric analysis to football is that the process isn't as systematic as the job processes those metrics can accurately evaluate. In football, You may gain 4+ yards against Jefferson and Lincoln HS, but Washington High is a different animal.
|
|
|
Post by fantom on Feb 1, 2015 15:45:51 GMT -6
If you think it helps you, and maybe it does, keep doing it. Fantom What do you suggest? Appreciate the feedback--the discussion helps me-- I guess Football, like education, has a ton of factors that bring or prevent success-- I know when our district started looking a some pieces of data we improved dramatically--it brought focus and clarity of purpose for everyone---were the factors of focus the only things that brought success---no but it brought us together to work toward a common goal and the results were positive. Are there other factors--yes--but without asking the question we would have never improved in the ways we have as a top school in the county.... I am trying to find what works with this line of thinking for football...and people like you have tons of experience and knowledge to share and exchange I appreciate your insight and assistance! I'm not saying anything other that what I said. If charting the plays works for you keep doing it. I don't do it but that may be because I'm a DC and have fewer calls to keep track of. It's not hard to remember the bad plays and, even in-game , see what went wrong. People work in different ways. If it works for you do it.
|
|