WE by count
Again using the Monte Carlo from the intentional walk tool and some data from baseball-reference on the league average OPS by count, I have cobbled together a WE chart by count. I am planning on using these numbers as the baseline for the intentional walk tool if the count isn't 0-0. But I thought people might be interested in this table stand alone so I have uploaded it here. Warning, it is a really huge file and probably will take several minutes to load on your computer.
A big thanks to John Walsh for pointing me to exactly the data I needed as sometimes this whole interweb is just too confusing.
Edit: Also a big thanks to tmapress and tangotiger for correcting my WPA WE error.
posted by Josh Kalk @ 8:50 PM
9 Comments
9 Comments:
Disclaimer: This is off-topic...
On the batter f/x cards there is a black box which seems to mark off the official strike zone. Aren't the pitch plots showing movement and NOT location? It seems like the strikezone on the picture is irrelevant, considering that a pitch outside of that black box could be a strike.
Dan,
Yes the pitcher cards show movement and shouldn't have the black box. The Batter cards though are showing final location not movement and should have the black box.
Am I understanding the pitcher plots correctly, that for say Joba Chamberlain, his fastball is breaking up 10 or more inches? And when you say if the ball was thrown without spin it would land in the original point (I can't remember your exact wording as I write this), do you mean it would land where zero and zero would meet? Awesome info here. Thanks.
Yes you are thinking about things exactly right azibuck. If the vertical break is +10 inches then the ball would have fallen another 10 inches if it was thrown without spin. If a pitch were thrown without spin it would be placed at x=0 and y=0 because it moved exactly like a ball without spin (because that is what it is). This has nothing to do with the actual strike zone as a ball with or without spin could be thrown at any location this is just the movement of the ball caused by the spin the pitcher puts on it.
I really like the idea of a IBB tool by count. However, I was wondering if you really need to utilize the WPA chart by count. When you enter the current batter's OPS into the tool you have now, you are really entering the OPS for that batter after the count is 0-0. e.g. top 7, runner on 2b, 1 out, +1. If R. Howard is up, his .976 OPS gives a -1.3 result. Say the count goes to 2-0. Howard has a 1.617 OPS after this count. So, wouldn't plugging in this new OPS, resulting in a +3.9, be the equivalent, and just as entirely accurate? All your MC should care about is what is this guy gonna do if I don't IBB him here. Does it really matter that the WPA decreased from 65.9 to 63.7? Or, should both this new OPS and new WAP be included? What do you think?
A small note:
WPA = win probability added = delta win expectancy of two points
WE = win expectancy = chance of winning the game
What is being discussed here is WE.
You are right Ben and that is what I am planning on doing I just have been swamped recently and haven't gotten too it yet.
tmapress, you are absolutely correct as well. I have WPA and WE interchanged before but that is no excuse for me to do it as well. I will try to go back and correct thing when I give the site a bit of a face lift this weekend.
I'm still confused about why you need to change your tool to incorporate the WE by count. Your WE of the orig situation (pitch or BB batter) is based on the WE and probability of all the possible situations after 4 batters. A change in count will change the batter's OPS, which only changes the probability of each situation, not the possible situations or WE of each.
In the moment before the decision is made, there is only 1 WE for that team. Both the final WEs after the decision and the net changes in WE would have the same difference between them. So I don't see how the league avg WE would help.
You could theoretically skip the batter's splits, and add the league avg improvement in WE from an 0-0 count to a B-S count to the WE you already calculate. Were you thinking along those lines? This should be very similar to the result if you modified the batters overall OPS by the league avg OPS count splits. But the individual splits, if statistically reliable, would lead to a more well-informed decision.
Ben,
The reason I want to add the count multiplier in is sometimes a manager will let a pitcher throw three or four regular pitches to a batter and then call for the intentional walk. If the count is say, 2-1 when the manager does this it is unfair to just use the regular tool because the batter now has a better chance of reaching base.
As for the splits you are absolutely correct. I could use a pitchers/batters true splits but the problem with that is I then can't use my linear approximation which is what makes this process so fast. So like you suggested, I am going to use a multiplier of what a league average hitter does on a 2-1 count divided by a 0-0 count. That ratio will then be multiplied by the users input to give the best guess as to the true OPS.
Post a Comment
<< Home