I'm no technological wunderkind, but I'm wondering if anybody would have the capability to compile data to chart the effect of pitcher fatigue. What I would love to see is if there is a linear increase in ERA as a game progresses (e.g. is a guy more effective in the first than the second; in the 2nd than the 3rd; etc.? Or, is there a threshold where he becomes less effective (e.g. a guy with 99 stamina throws his 60th pitch as effectively as is 1st, but by pitch 80(???) he starts to decrease in effectiveness.

Does anybody have the data gathering capacity for such a project? If Bill James played HBD, he would thank you.

You will have to settle for the pittance of a consolation prize that is my very own meager gratitude, instead.

Bless you.

The biggest problem is we don't have that data. While we have overall pitch count data and batters faced data, we are not given the amount of pitches for each batter or even each inning.

Probably the best one could do would be an OPS for each batter faced relative to stamina.

Posted by topoftheworl on 12/17/2016 10:46:00 AM (view original):
The biggest problem is we don't have that data. While we have overall pitch count data and batters faced data, we are not given the amount of pitches for each batter or even each inning.

Probably the best one could do would be an OPS for each batter faced relative to stamina.

But even then we don't get in-game fatigue/stamina numbers so the OPS against wouldn't be all that telling?

I think you could do OPS against per inning but you'd have to dive into each AB against that pitcher per inning and guesstimate how many pitches were thrown for each IP.


And even if you did that, I feel like it wouldn't be conclusive.

at minimum, i'm just saying take a starter's era in the 1st inning. compare it to his era in the 4th inning. 99% of starts go into at least the 4th, tandems excluded, i'm guessing. i worked at a tech company full of people who could do things i couldn't believe (like grow beards that ran from their cheeks to their chests). just checking if this is possible.

It could certainly be hand compiled, but you couldn't get enough data to produce a result with any power.

i lack the technical prowess to write something that could scrape the data. If someone could I could process the data.

I actually work in digital analytics, but I think there are entirely too many variables to do anything that makes sense. You could base it on overall pitch count and then examine where each pitcher allowed runs in the game, but each team / lineup / batter has ratings that come with their own probability modifiers. The only way this would work is if you had a large enough sample size of one pitcher facing a lineup full of batters with identical ratings over and over.

Easy/unfair/possibly irrelevant answer - be conservative with pitch counts, and spend money on a solid bullpen.

My gut instinct (so take it with a grain of salt) based on box score observations is pitcher fatigue level does not affect effectiveness unless he where to hit 0. I can't count the number of times a guy gets lit up in the first only to cruise the remainder of the the game and meet his pitch count.

Can't speak to pitchers who enter the game fatigued as that's not something I make a habit of doing. Have to ask the tankers, they'd have a better idea, if they even paid attention to the boxscores.