Yep, and forget all the other times running the ball worked in a similar situation.
The math behind going for 2 when down by 14 late in game and scoring a TD.
- Thread starter Grant Green
- Start date
Big difference there is you must win game six to have any chance of winning the series (as opposed to “going for two” situation).
seems more akin to the nascar example where an intermediate result has value apart from the final result.
But what if I really wanted the goat. Maybe I'm starting a farm or have a big patch of poison ivy I need to clean up?
Is the percent of winning while not switching always 1÷n? Therfore the percentage of winning while switching is 2×(1÷n). Where n is the number of doors.
So, more doors makes switching less beneficial. Still twice as probable as staying, but the magnitude is smaller.
analytics assumes all opponents are the same. They never account for opponents having a stud player out or coming back from injury. It generalizes such that playing Rutgers will have the same effect as playing Georgia.
But you’d also need to win game 7. You should make decisions in trying to win game 6 that hurt your overall chance of winning if the end goal is to win the series.
I'm still not sure how some argue against using analytics. No you don't do it 100% of the time, but you can use it as a guide. I'd much rather have a HC that uses current data and methods then a dinosaur who isn't willing to adjust. This back in my day we just used our gut is like saying why break down film....just out execute everyone. No need to try and get a leg up legally. There are very few businesses and even pro sports that don't take analytics into account. It's not a black and white thing, but yes.....coaches need to have that data available to them to make an INFORMED decision.
It would depend on how it goes down. If there are more doors and they just open one, the chances of winning aren’t as high (though still better to switch). If they keep opening until you’re down to 2 doors, then chances of winning are even higher with the switch.
I had a game theory class my last semester that covered some of these types of things that I really enjoyed.
Last edited:
That’s why you’d use analytics to help guide your decisions, not take one set of data and use it as gospel (plus, it’s possible to separate out the data and look at opponents differently…we don’t have 100 people on our football payroll for nothing).
Thanks that makes sense...
but I would rather roll first and then choose. My chances of winning would be much higher.
Kind of like logic has no place on a fan message board.
It is important for a coach and his assistants to know the analytics of any given situation. But whenever I see a coach go for 2 points before the 4th quarter because “the book” says to do it, it is cringeworthy. Teams routinely regret chasing points too early in a game.
Why not? If it helps you win more games, why would you not use it?
Math is applicable to many parts of ordinary life.
I find the term 'chasing points' to be cringe worthy. More sports dogma that has never been proven to be true. If an action gives you a higher chance of success over the long run, you should probably do it. Saying that teams routinely regret 'chasing points' is completely anecdotal and you are likely only remembering the times it fails and not the times it works.
Is the math generated from 10 lesser teams equal the math on would get with tOSU and Michigan?
So you should play to lose against better teams or ignore the math because it scares some old people afraid of any change?
Show me how the math changes when, during the game, the opponents starting cornerback left for injury and his replacement just hobbled off the field.
When you running back just had two consecutive runs of 15 and 20 yards, does the math say to keep running it in that drive until they stop it, or to start throwing the ball instead?
So in your world a HC should never use trends or tools available to him.....got it. Why should coaches break down film anymore? Why would they want the data to know what an opponent may or may not do? You just need to go out and execute, right. Who cares if a player gets hurt or not, just execute.
Analytics have been in football for a very long time, but now they are getting into the game even deeper. Team A runs the ball 70% of the time in this situation in this part of the field....so just ignore that information as math is bad. You realize that data was being used for decades now and it's more available with the better BI tools that are out there. I can guarantee you every major university is using analytics in a huge way in more than just football.
Not to mention this is about when to go for 2, punt, and you're now breaking it down to every single play.....that isn't what this is about but I can see how these things confuse you so easily. As far as changing from run to pass, maybe the QB's and OC should ignore how the defense lines up too. Just make a single call and stick with it as doing things that may work in your favor based on data and the past is voodoo. Any HC worth his salt isn't ignoring analytics anymore, but doesn't mean they only rely on that to make decisions. I know in your world everything is black and white, but this topic is a bit more nuanced for you IMO.
Last edited:
That's a different discussion. You are referring to the quality of the data, which doesn't prove the blanket statement that math doesn't belong in football.
But, if "chasing points" doesn't exist, shouldn't every team go for 2 every time? Maybe it depends on what you mean by "chasing points". For me it means after doing something with a lesser chance of success but a potentially bigger payoff, if you aren't successful, you need to adapt your strategy to account for the smaller payoff you didn't obtain at a higher probability for success. So, in my estimation, going for 2 early in the game leads to the potential of "chasing points" but later in the game (typically 4th quarter) that is less of a factor as you number of future possessions approach zero.
It depends on your probability of success. If your XP% is 94% and your long term average 2pt% is 48%, technically you should go for 2 every time. Over 100 attempts the XP yields 94 pts and the 2pt try yields 96pts. Will any coach do that outside of highschool? Probably not. If your 2pt% is poor, then yeah, maybe better to get the XP, but that is at least better reasoning than "well, I heard that it's always bad to chase points".
Here's another angle. What about when Chip Kelly and Mike Tomlin would go for 2 on the first TD of the game. Perhaps they felt that they had a higher probability for success early in the game? If you make it, now the other team is "chasing points". I think when people analyze these situations they tend to only think of the bad outcome and it's effects (my team will fall further behind) but not consider the good outcome (it could end up winning me 1 extra game this season which could be the difference between playoffs and golf).
I agree that it is possible that it could win you that extra game. But it is probably more likely that it could lose you that extra game which also could be the difference between playoffs and golf. Again, otherwise every team would do it all the time.
Why is it more likely to lose? You can't make that blanket statement. If a team is awful at converting 2pts than, I agree they will lose more. If a team is really good at it, they are more likely to win. For example, if the Eagles could make the tush push work an average of 50%, they should do it every time.
The reason that coaches don't always make the statistically correct decision is often due to public perception and job preservation. If the Eagles started going for two every time, Nick S would get eviscerated the first time they lost because of it. As I said earlier, Malcolm Gladwell does a great podcast segment on this - check it out in link below. He follows it with commentary on the movie No Good Deed which is really interesting too.
I agree that sometimes coaches don't do the right thing. I guess that I need to better understand where you fall on this. When you say that chasing points doesn't exist, to me it sounds like you are advocating going for 2 early if not always because it will pay off for you a certain percentage of the time. It's hard to say exactly how often without exact numbers and, as many here have said, individual circumstances will vary. If that success rate is less than half of the XP rate then it seems that it would be more likely not to be beneficial.
I guess my issue with using math in these situations is that using averages to calculate one play. What happened in the months of September and October might not be relevant to a team in January. Analytics try to make a simple math decision, but the number of variables make it a much difference calculation than using a simple average.
And that should play into those things that are tough decisions and/or right on the edge...do I go for it on this 4th and 2 from the 50? When it comes to the initial question in this thread, the decision should be a no-brainer under pretty much any scenario. When you're down 14 and then score a TD, you should go for 2 in pretty much any situation with a certain amount of time left on the clock.
Another one that might start popping up is how to handle things when you're down 10 and score a TD...it is situation dependent to a degree, but there is a very strong argument to go for 2 in that scenario. I'm sure that a coach doing so tomorrow would blow people's minds.
Probably if you look at the math it never makes sense to kick an extra point or punt. You should probably "always" go for it.
That's not correct...the math will often tell you to kick an XP or punt, it's just way less that coaches have traditionally done.
I think a message board filled with Mr Spocks would be a dull place.
What irks me about people using the term "chasing points" is that 1)they tend to use it every time a coach is losing and goes for 2 and it isn't near the end of the game, 2) believe that it is always a bad move because they heard someone say it on a broadcast when they were young. I would definitely agree with your last statement (unless late in the game and you have to gamble). And yeah, if your teams 2pt rate is better than half your XP rate, technically you should go for it every time. I understand the psychology of a coach not doing it though (back to Malcolm Gladwell - check it out, I think you would like it).
how do you determine your teams 2pt rate which, unlike an XP rate, will vary depending on who you play
I'm certainly not an expert, so I can only give a rough explanation. I believe they will mainly use in-season stats for 3rd and 4th short yardage success rates of both your team and the opponent. Obviously, previous 2pt attempts by both teams fill factor in. Probably a small contribution of league average. I'm sure there are more factors that guys smarter than me have figured out.
League/historical average should be a big part of it, since the sample sizes for isolating just yourselves or your opponents would lead to a pretty high variance.
No exception to that, especially for NFL. Early in the season, probably have to use previous season data and understand that it won't be as reliable.
Even at home? After scoring on two recent, if not consecutive drives? With players amped up and the other team trying to recover? Yeah, run a low percentage play for two and deflate it all.
Why are you assuming you’d run a low percentage play? Yes, even at home after scoring on two recent drives, going for 2 when down 8 should be a no brainer.
The whole "emotional letdown/momentum killer of missing the 2pt conv" angle is silly to me. I wouldn't factor it into the probability at all. Even if it were a factor, it wouldn't be remotely close to overcoming the ~12.5% edge you are getting.
@Grant Green - are you into the sabremetrics side of baseball at all? I’m not, but worked with some of Voros McCracken’s stuff on the soccer side years ago, and he was the one that came up with the analysis around DIPS…those findings blew my mind even being very deep into the statistical/modeling side of sports.
