Quote Originally Posted by tallguyindc:
It should be closer to 35%. The math works at anything above about 14%.
https://www.advancednflstats.com/2009/09/4th-down-study-part-3.html
I really don't want to get into a general fourth down discussion but I agree with him. You're right the math says that you should generally go for it on 4th and 1 and that you need to apply game-specific factors. But if on average you should go for it and coaches apply game specifics and never go for it you are quickly approaching Lake Woebegone territory where they proudly boasted that "every child is above the town average"
Those mathmatical formula's give a very rough idea of the odds of each senerio, but are not accurate.
It seems you read that on a web site, it sounded good to you so you bought into it without alot of further thought.
First, he goes off recent historical averages, which "DO NOT" account for the skill and ability of each opponent in this game. It's flawed already
Second, and the "biggest flaw", is teams "do not" perform at a average level each game, some games they perform well above their avearages, these are games they normally win, then they have games of performing well below their averages, these are games they might lose, that's how stats go in the real world.
Let's look at Dallas's average 3rd down percentage, then let's break it down to a game to game and see how many games Dallas performed at their average.
I use 3rd down's because it's a larger sample size than 4th down plus, if teams chose to go for it on 4th downs every time as the web site states would be the right play, then they would have a much larger sample size on 4th downs very similar to 3rd downs now.
Dallas season average 39.7 %.
Game 1 --- 38%
2. --- 46%
3.--- 50%
4. --- 50%
5. --- 52%
6. --- 0%
As you can see by the info above Dallas played at it's average just 1 game out of 6, and they played extremely off their average in 4 of 6 games. This is how stats work in the real world.
Now, how then can putting in a historical average give you an accurate odds of Dallas on "this day" ????
Simply not possible !!!
For that mathmatical formula to be accurate, each team would have to perform at an average level in each situation that makes up the formula, and as the info above clearly without any question shows, it does not work that way in the real world.
In other words, in games that a team is playing above their averages, these are games the team is highly likely to win, when a team is playing above it's average that would increase it's odds in the formula if you inserted these above averages into the formula which would be representive of how the team was playing on this day.
And it would be the opposite when teams are playing below the averages on a given day.
Human beings playing a football game are not like a coin, where regardless of how many times a coin comes up heads, it's still 50-50 it comes up heads on the next flip. Coins not posses skill, statergy, heart or desire and motivation.
A football game involves skill and ability, coaching, energy, heart, motivation and desire on this given day, all those things come together at different levels each game, to "assume" each team plays at a historical average each and every game is silly.