Gridiron Freakonomics
This is fantastic. What happens when you hand an accounting major a whistle and make him a high school football coach? He goes all Moneyball on you and changes the way the game is played entirely. Enter Kevin Kelley, head coach of the Pulaski Academy Bruins. Thanks to his devotion to Malcolm Gladwell, Freakonomics, and all things statistical, he has changed the face of football in Arkansas. How? He never punts. Never. Goes for it on every fourth down situation, even on his own goal line. He also kicks an onside kick every time he has to kick the ball off. Sound crazy? Watch this to see the mathematical reasoning behind the decisions, and the amazing results he's attained by sticking to it. Then try to imagine preparing to play against him.
Fantastic. Thanks Eddie
They must have a stellar defense.
Awesome. Wonder how different the numbers would be were these tactics applied to the NFL or even college
I think the skill of the kickers/punter plays a huge role in it. If you watch the video, the opponents' average starting spot on regular kick off was the 33 yd line - on failed onside kicks it was only the 47 yd line. In the NFL, the starting point on regular kickoffs is probably much closer to 20. The change in field position on a punt is also drastically different in high school (probably ~35 yds) vs. the NFL (probably ~47 yds, with better return coverage).
In summary, the stats probably don't work or don't work as well at the professional level.
To sum up what you said, the strength of defense related to offense is much closer at higher levels of play compared to HS which is much more offense-heavy, generally.
The biggest problem was that, as mentioned by "accountingbyday" below, no pro coach would want to take the risk in front of the public, even if statistically it made sense to.
Agreed with this 110%. I would add that "money ball" would probably work in college and the NFL, but you'd use statistics to make different decisions.
Statistics are going to be changing football.
You won't see anything to this extreme in the NFL for a long time (maybe ever) because all it takes is for the "new" philosophy to not work at the wrong time and the coach loses his job. It's not worth it for them.
Mark Trestman of the Bears in one who is starting to bring in a new level of thinking. At the end of Sunday's game as the Ravens were close to a potential go ahead score Trestman did not use his timeouts. Everyone thought this was the wrong move bc Bal could score a touchdown with 10-20 seconds left and the game would be all but over. His explanation afterwards was fantastic and based in statistics.
I am a Bears fan and I was starting to freak out at him not using those timeouts. The way it played out in the Ravens game was absolutely perfect and essentially pigeonholed them into throwing on 3rd down whereas if he had used his timeouts they would have both options (due to them saving their timeouts). That took serious balls and regardless of outcome I prefer it to Lovie Smith any day of the week.
I'm a Bears fan as well and had a similar reaction. Obviously, I was happy that they ended up winning, but I didn't like the decision at the time. Once I heard his explanation I changed my tune.
I completely agree about the preference of Trestman to Lovie. Trestman gave more thought in that situation and answered the question with more insight than Lovie did in the entirety of his time in Chicago.
I still question why Jay played so long 2 weeks ago and being thoughtful doesn't in and of itself make a successful coach, but I definitely appreciate Trestman's approach.
That's why I always go for it in Madden.
I wonder how long before he gets called up to the NFL? Some struggling team is going to want to try this, yes?
I feel like anyone who reads sports blogs has discussed stuff like this ad nauseum. Why kick an extra point when you can go for two everytime, etc. Its a cool idea but it will never be done at levels where a persons major source of a huge income is coaching as there is just too much on the line.
what if you get a forward thinking owner? To say it will never be done I think is an overstatement. Look at the innovation coming to the NFL this year with the Eagle's offense....there is evolution in every game over time as new optimal strategies are uncovered over time and as rule changes shift (ie passing has become much preferred with all the protection afforded to recievers now)...
Youre also talking about a sport that is literally owned by conservatism. A coach goes for it on 4th and short two games in a row and hes the most aggressive guy on the planet. Obviously, saying never was hyperbole but in high school the delta in talent is astronomical whereas in the NFL there is enormous parity. Further, you have a high school shop teacher coaching football for an extra 5k a year not a guy who has worked his entire life to get a job as a head coach making 5MM. Its not exactly apples to apples.
Imagine the mental edge over your opponent that you bring into the game when you know you've got four downs every time. Add that to fundamental disrespect you show to their special teams squad by kicking an onside every time. Just a complete mindfuck every time your team takes the field.
not sure what makes football so much immune to this than baseball.
was wondering about why not at least punt at your own endzone but he makes a good point, 92% of time other team will score if you don't convert on 4th, and if you punt they'll score 77% of the time because of good field position, not a huge difference, especially considering on average they're converting 50% of 4th downs, great # for a high school team, most nfl teams are in the 40% range i believe
This sort of philosophy is already making headway in the NFL, as more coaches elect to go for the conversion on 4th down in lieu of punts. The data are unequivocal, even at the NFL level: going for it on fourth down is almost always the best decision.
Here's a much more consumable (and shorter) reproduction of the study linked above:
http://emlab.berkeley.edu/~dromer/papers/JPE_April06.pdf
Moreover, studies also demonstrate that NFL coaches call far too many run plays (see below). Indeed, the play calling is, overall, a great deal too conservative at all levels of American football.
http://www.tangotiger.net/files/kovash_levitt_w15347.pdf
Pretty huge difference between going for it on 4th and goal on the two in the first quarter and going for it on every 4th down. Optimal strategy is obviously somewhere in the middle.
Only in the loosest of senses. On almost every part of the field, it is worth going for it on every fourth down of fewer than 4.5 yards-to-go. Only inside your own 20 (where going for it is still economical with 3-or-fewer yards-to-go) and at your opponent's 20 (where going for it is only economical at 4-or-fewer yards-to-go) does that idea change. Between your opponent's 40 and 30 yard lines, going for it on anything less than 4th and 9 is the optimal decision!
Perhaps more importantly, the data show that coaches universally and consistently make the wrong decision!
I'm all for evolution of the sport, if it's not of the variety that get a BS penalty called on Ahmad Brooks for rocking your boy Drew Brees.
It makes the most sense when you're on the opponent's 40 yard line when it's fourth and short. A punt will likely only net you 20 yards. Riverboat Ron and the Panthers are among the best at this.
Side note - Bill Barnwell on Grantland is the best in the business and talks about this kind of stuff daily.
http://www.vegaswatch.org/2013/09/how-not-to-handicap-prop-bet-2013.html
The guy is great with in game type work but his sports betting stuff isnt his strong suit.
This thing works great in HS football because the talent and/or experience are not fully developed.
The only thing I think would work would be to go on 4th down 50% of the time. If you go all the time, teams will know and will adapt to you.
On-side kick won't work in the NFL: 20% conversion rate would plummet to 5-10%, and more importantly in NFL with a good kicker (Patriots, Raiders), you can have a touch back all the time, forcing the opposing team to start at its own 20.
Agree. The data on onside kicks in the NFL are much less persuasive.
http://www.advancednflstats.com/2009/09/onside-kicks.html
Like many, I'm curious about the stats he provides. Specifically, I'm wondering about those stats that say your opponent has a 92% chance of scoring if they're at your 5, and a 77% chance of scoring if they're at your 40. Scoring can mean a field goal or a touchdown, but I think it's important to know the breakdowns of the type of scoring. Unless you have Alabama's defense you're likely not going to shut anyone out, but if you can keep the other team from scoring TDs and limit them to only kicking FGs, while you score TDs, then you have a great shot at winning.
That being said, the context is important here too. Such as, how good is his team relative to the teams they play? If they're near the top, then it's not as important what you do on 4th down. On the other hand, if your team rates near the bottom of the "league" you're in, then you probably don't want to adopt this approach.
gotta agree with HappyPantsMcgee on this one...optimal is happy median
This is a golden mean logical error. In this case, the "happy median" is much further in the direction of coach aggressiveness than is currently implemented.
Like the innovative thinking but you would have to make some adjustments at the NFL level. In addition to high school vs. NFL punting disparities, one thing to consider is the difference between field goal kicker quality in high school vs. the NFL. In high school, anything beyond PAT distance is a toss-up, whereas at the pro level once a team gets to the opposing 40 yard line a field goal is a very real threat. Therefore, turning over the ball anywhere on your side of the 50 yard line would almost guarantee at least 3 points for the other team every time. Seems like it might be hard to overcome that kind of deficit.
And yeah, onside kicks succeed almost never in the NFL so again, you'd be be giving your opponent 3 points on every kickoff.
http://emlab.berkeley.edu/~dromer/papers/JPE_April06.pdf
Almost never succeed?? Try again. I don't think anyone would advocate consistently taking onside kicks in the NFL, but it's much less decisive than you're making it sound. Onside kicks, at the very least, appear to be underutilized.
http://www.advancednflstats.com/2009/09/onside-kicks.html
@Northsider notes: "The only counterargument available is that coaches perplexingly pursue sub-optimal 3rd down defensive coverage that would be improved (somehow?) when facing 4th down conversions. Dubious, at best." (Sorry can't get quote function to work on my laptop)
I don't necessarily agree with this. (I haven't read the study, btw). I think what the coaches consider "optimal" playcalling on bothO and D will be different on 4th downs when you have to either convert or turnover the ball. Not saying that invalidates your points. But one might argue that the playbook gets smaller on 4th down, which could change things.
Of course, if you're going against my beloved Notre Dame, they'll run an empty backfield and call some x-box pass play with Tommy Rees, and the D has the advantage before the ball is snapped.
I mean, at best this has nominal effects. The study provided would merely suggest that the offensive coach change nothing from his typical playbook and continue as if 4th were "just another down".
Perhaps defensive strategy would change, but if there were a way to meaningfully reduce the E(V) of a 4th down conversion attempt, it immediately begs the question: why aren't they applying this defense on 3rd down? I see no convincing argument as to why we should expect such a difference in the outcomes.
This thread filled up fast. But I did want to mention that advanced statistics are changing the face of all of the sports.
3 pointers in basketball didin't exist initially; were used mainly in comeback situations in the 80's until guys like Larry Bird started using them when they were open in the 90's; turned into a mainstream strategy in the 00's; and are increasing drastically every year in the 10's. Statistically, despite the lower FG% from 3PT range, you still maximize your points per possession by playing inside out and going for easy layups or just launching 3s. Over the course of a game with an average of 100 possessions, launching 3's at a 35% rate is theoretically more effective than getting 2's at a 50% rate.
Moneyball is a thing, and not just in baseball. I hope some NFL team tries a no-punt strategy. I haven't read the study yet, but if the numbers suggest that it could work, I wouldn't necessarily bet against it. Some "advanced"statistics are bunk, but many counter-intuitive (read: non-traditional) methods are being proven by statistics. In basketball, my sport of choice, this has certainly been the case.
@Northsider I agree with your points. And I think 4th down decisions in general are an interesting topic for statistical analysis (again, I haven't read the study) because this seems to be one of the biggest "go with your gut" calls for coaches.
Even if you accept that the no punting / always onside kick approach yields a statistically significant difference in the outcome of the game, the increase in volatility caused by always going for it on 4th down makes doing this virtual career suicide in the NFL. Just as easily that the team goes up 3 touchdowns in the 1st quarter, the team could be down 3 touchdowns and what then? Are you still going to go for it on 4th down at your own 20 yard line?
Additionally, these statistical advantages are only applicable if you think the two teams are equal. If Jacksonville were to implement this system, I would expect that it would have a negative expected value as 4 downs for jacksonville means more chances for turnovers in addition to the higher probability that they turn the ball over on downs. The only teams that could implement this system are the elite teams (GB, NE, NO, DEN, etc) and they are already winning so what's the advantage to taking on additional risk?
At the end of the day, I would want to look at the volatility adjusted expected value before considering this approach.
Didnt you read the rest of the thread? If you dont think this makes total and complete sense than you are bad at math and dont know anything.
very good point on volatility -- an important consideration when the outcome (win or lose) is binary. For more elite teams, against inferior opponents, this might increase the expected margin of victory but increase risk.
I would argue that for a truly inferior team, wouldn't you want to take a higher volatility/risky approach?
I was just about to say that. One would think that the bottom 10 teams could benefit by taking on the risk.
I think a deeper level of statistical analysis would need to be utilized for each team. I agree with glide that some teams would not benefit by taking on greater risk. However, others may only benefit by taking on more calculated statistical risks that are in-line with their level of talent. Maybe a team with a putrid offense will only go for it on 4th down when they are past the 50 yard line, for example.
Precisely what the strategy advocates suggest.
I mean, obviously it poses career risks. If it didn't, rationally, all coaches would be implementing this scheme, given that it maximizes win probability. As for your anecdote, the data indicate that you'd be far more likely to be up by 3 touchdowns in the first quarter than down by 3 touchdowns after going for it on 4th down more often vis a vis conventional strategy.
This is nonsense. These statistics are collected across multiple seasons and games involving both balanced and imbalanced matchups. There are no implicit assumptions re: "equal teams".
Teams cannot utilize a posteriori knowledge to retrospectively rationalize a priori decisions like you can 11 weeks into the season. Given that teams don't know their "power rankings" ahead of time, they must make decisions based on a theoretical coaching framework that applies optimal strategy. Why it is that you believe that coaches should apply a strategy that, on balance, reduces win probability is beyond me.
Even presuming a clear way to determine talent mismatches, the conclusion is the opposite of the one you suggest. Contrary to your assertions here, the primary advocates of this strategy are pushing for poorly performing teams to implement the strategy, since they have very little to lose from high variance strategies. Pursuing conventional strategy consigns them to a sub-par win/loss record; why not pursue a high variance strategy that will (say) either end with you going 13-3 or 3-13, rather than taking the beaten path to a 7-9 playoff-excluding record?
Then read the study. What relevance does the variance of score have to optimal strategy? Win/loss outcomes are binary; it doesn't matter if you lose 65-0 or 65-64. The only question is: what is the strategy that maximizes your win probability? And the data are clear.
If you're skeptical, implement on the 2-sigma band to ensure you're getting the right "risk-adjusted returns". No matter how you cut the numbers, 4th down conversion attempts are dramatically underutilized.
This is beautiful, similar to financial markets, he is exploiting the imperfections of the game...
Another issue is that football coaches are of notoriously low intellect. I can't believe how bad they are at managing clock at the end of halves. Giving the average head coach the tool of statistics may prove of no benefit if they don't know how to interpret the numbers.
I remember reading something on this a while back. While I admit watching a game like this seems exciting, I do not see widespread adoption anytime soon. Partly due to the ability to lose your job in only 1 volatile season, and partly due to the slower nature of strategy adoption in the NFL.
I would even say that adoption would take place at the college level first and then the NFL. The resurgence of the spread offense(truly an older type of football started in the 30's) is probably the most poignant example of this.
I honestly look forward to the day that statistics is used heavily in the NFL for optimal strategy. Definitely for selfish fantasy football reasons, but also for added excitement every offensive drive.
@Eddie, I agree it would be a total mindfck to have a team tell you their 4th down and kickoff plan for the whole game. Probably on the level of stabbing yourself before a knife fight, and then yelling "bananas, monkey shit,your mom!"
Great post! Although, I'm not sure how this would fare at higher levels, but at the high-school level, it's brilliant. It reminds me of the offense that the Univ. of Oregon runs (call it gimmicky/scam if you'd like), but Chip Kelly has numbers behind that offense.
I think everyone agrees that going for it in certain situations on 4th down makes sense. But, the idea that it should be done anywhere near as frequently as some bullshit highschool team does it is madness. It really makes sense when you're going to leave your opponent in bad field position even if you don't get the first down.
As for the onside kicks. Doing that every time would render them useless in the pros. The idea that using them out of the blue ever so often makes sense as a means of catching a team off-guard. Using them all the time just doesn't work.
agree with the onside kick thing, especially since in the NFL, the kickoffs result in touchbacks much more frequently now. (so field position differential more significant and the success % dramatically drops once it's expected as was noted before)...but still, could be used in say ~1/10 kickoffs by some of the shitty teams to increase the volatility / variance of outcomes and still maintain a high enough success % to make it a positive expected outcome...
As far as 4th down is concerned, however, I think the studies cited in here point to an interesting statistic showing very compelling evidence that NFL coaches tend to go for it on 4th down at a rate that is WAY too low to yield the highest % chance of winning the game...
Did someone actually throw monkey shit at me? Did I say something controversial?
LOL. I know the feeling.
Unde quis repudiandae praesentium accusamus. Fuga omnis dolorem quo nihil. Voluptatibus vero sunt quae illo dignissimos nesciunt facere. Sed id omnis veniam nulla minus omnis. Officia cumque fugiat velit tempora doloremque tempora.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...
Corrupti et totam quia et. Quae ut esse eaque labore aut fugiat autem.
Neque iusto ea a omnis ullam odit. Earum quas id porro delectus necessitatibus qui. Aperiam incidunt consequatur rerum labore quia doloremque debitis temporibus. Autem voluptates odit quo id animi dolor sit. Voluptatem quos aut consequatur sapiente harum ratione dolores. Rerum ut amet autem reprehenderit distinctio veniam.
Aut in et sit. Nobis enim sed libero aliquid aut quia recusandae. Praesentium incidunt aut quia repellendus beatae magni officiis. Nobis a qui dolorem mollitia. Modi eum perspiciatis distinctio.
Delectus voluptas iste voluptatem est quas eveniet. Excepturi ut et facere dolorum recusandae omnis atque. Quasi sint commodi quis repellat soluta quas et. Corrupti odit aliquam labore natus rerum corrupti optio ratione.
Accusantium officia vel ipsa at rerum mollitia necessitatibus. Temporibus molestias ut qui nesciunt. Iure nostrum enim a quia suscipit. Voluptate eos explicabo autem eos.
Rerum consequatur et ratione quia. Quibusdam est est temporibus porro dolorem autem nemo. Et omnis autem non deserunt laborum corporis tempora ab.