Bastardized Pitch Edits 2015

[ilstu24]

Re: Bastardized Pitch Edits 2015

So whats the status on these? The edits are done and just waiting to be input?

[QuestGAV]

Re: Bastardized Pitch Edits 2015

So whats the status on these? The edits are done and just waiting to be input?

First pass of edits are done other than FAs. I think there are only a handful of them worth editing anyway. There's a buttload of guys I want to create from scratch that weren't on last year's sheet, that'll take a bit of time. I'm also working on setting it up so that anyone can copy & paste in the Brooks data and get a full set of ratings for a guy going forward - this will take some time too but be pretty damn useful in the long term.

[QuestGAV]

Re: Bastardized Pitch Edits 2015

Okay, massive update today to the sheet to make it more usable & sustainable long-term. It really took looking at several thousand pitches multiple times to be comfortable coming up with all these formulas. The players on the sheet from seanjeezy's 14 sheet are largely not rated by these formulas but the newly created players for '15 will be.

COMMAND

For starters, I came up with a formula I like for control after meddling around with an idea that Ohiocub came up with.

Command = 85+usage%-(ball%/10)-(LD%*2)-(HR%*40)

For an example of how it looks across the repertoires of a few guys look at the 4th sheet labeled "Control". The result is in the "Command" column at the end of the raw numbers from Brooks. Basically it assumes that a guy who uses a pitch often has a fair amount of confidence he can put it where he wants to. He then loses credit for those 3 bad outcomes. It's not weighted quite as heavily towards HR% as it looks because that raw number is much lower.

It makes more sense when you look at the results for the pitchers on that sheet. In general the formula seems to yield a fairly good range of numbers from 0-100 and hopefully encourages pitchers to throw their most used pitches more often in the game.

MOVEMENT

I also now have solidified formulas for both classifying pitches based on their movement and rating them in the game based on that movement.

In the leftmost column in green you will see the basic pitch types noted in Brooks. Underneath in the same column you see the pitch types in the game and in the 2nd column the movement criteria for that pitch. In the 3rd column is the movement based formula for that pitch.

For a few examples, a "fourseam" in Brooks is always a 4SFB in the game now. The movement rating for a 4SFB is Hmov*10.

It gets a little more complicated if trying to classify any Brooks Splitter or Changeup. For starters, any changeup or splitter with > 7" Hmov and > 7" Vmov becomes a forkball in game. You'd then apply the forkball movement formula: (Hmov*8)+(10-Vmov)*2, emphasizing the lateral/horizontal movement that is more apparent on forkballs in the game..

Any Brooks Splitter that doesn't meet the FORK criteria stays a splitter in game. Any changeup with > 6" Hmov becomes a circle change (or if they're known to throw a circle). The rest become regular changeups. All splitters and changeups (not including forkballs) use this formula for movement: (Hmov*3)+(10-Vmov)*7. This formula emphasizes the vertical movement that is more apparent on splits/changes in the game.

SUSTAINABILITY

I'm relatively sure none of this makes any sense if you're not actually looking at the sheet as you go. It's not entirely necessary to understand where each of the numbers come from if you can follow the directions to classify a pitch, as the other new addition to the sheet is the ability to copy & paste two rows from Brooks into the sheet to get the pitch type, velocity, command rating, & move rating.

Start by copying and pasting in the "Trajectory & Movement" and "Pitch Outcomes" tabular data for each pitcher at the top of the "Classification" sheet. Select the line of the pitch you want ratings for. Classify the pitch based on the Hmov and Vmov. Copy & paste the line for that pitch in the line below the classification & movement formula. The line below the data you pasted in should change to the ratings for the pitch you want.

Obviously the sheet I have is locked so you would have to make a local copy in order to edit the data. But this should give anyone the ability to come up with a set of reasonable ratings for a pitch without having to reinvent the wheel on a yearly basis. Or update ratings as the season goes on. Or add new pitchers as they come into the league/generate enough data for a reasonable set.

[Bravescountry123]

Re: Bastardized Pitch Edits 2015

Some are in yellow which means means they are being worked on? Also low drop changes to forks are left right?

[PhilliesFan13]

Re: Bastardized Pitch Edits 2015

Haven't seen the sheet yet as I just got home, but does it no longer contain the numbers we would be entering in for the pitches like before? We have to use the formula to calculate it?

[QuestGAV]

Re: Bastardized Pitch Edits 2015

Some are in yellow which means means they are being worked on? Also low drop changes to forks are left right?

Correct.

Haven't seen the sheet yet as I just got home, but does it no longer contain the numbers we would be entering in for the pitches like before? We have to use the formula to calculate it?

No, it still has the numbers to enter. It just now has the ability to generate the numbers as well. If you feed it some yummy data it will poop out a set of pitch ratings.

[SFCFeagin]

Re: Bastardized Pitch Edits 2015

This question has prob been asked and I may not be seeing it. But why do sites such as Fan Graphs show different pitch types and velocity than Brooks? I mean they seem close in general, but sometimes it can be 2 types of pitches that are omitted from one site to the next.

[OhioCub]

Re: Bastardized Pitch Edits 2015

This question has prob been asked and I may not be seeing it. But why do sites such as Fan Graphs show different pitch types and velocity than Brooks? I mean they seem close in general, but sometimes it can be 2 types of pitches that are omitted from one site to the next.

The other sites realy on a computer to generate the pitch types but Brooks uses a computer to generate the raw data and then determines pitch types through actual human interpretation. At least that's how I've come to understand it.

[QuestGAV]

Re: Bastardized Pitch Edits 2015

Yeah, they each get raw data and use a variety of algorithms to classify them. Brooks claims to go a step further in terms of manually verifying the pitches and the data seems to back that up. Without that 2nd step, guys with inconsistent release points will sometimes seem to have an extra pitch. The scatter charts for release point are really useful in cases like that.

[55]

Re: Bastardized Pitch Edits 2015

For the guys who have an empty box for their command ratings, how are you entering them on the roster?

[QuestGAV]

Re: Bastardized Pitch Edits 2015

For the guys who have an empty box for their command ratings, how are you entering them on the roster?

Leave them as is if it's a pitch the guy already had. If it's a new pitch use OVR for 1st pitch, OVR-20 for 2nd, OVR-30 for 3rd, etc. If you feel like going all out you could put the Brooks data into the sheet and see what it puts out.

[55]

Re: Bastardized Pitch Edits 2015

Leave them as is if it's a pitch the guy already had. If it's a new pitch use OVR for 1st pitch, OVR-20 for 2nd, OVR-30 for 3rd, etc. If you feel like going all out you could put the Brooks data into the sheet and see what it puts out.

Thanks, man. I'm going to leave as is if it is an existing pitch and then just do the OVR, -20, -30, etc. when I punch these into the OSFM v2. If I start messing around with the sheet too much, it'll give me bright ideas that I shouldn't get!

Data entry only for this guy.

[rjackson]

Re: Bastardized Pitch Edits 2015

I think there probably should be some small correlation between usage % in real life and command rating in game... ie a guy throws a pitch more often because he knows he can put it where he wants to. But I think making that the only factor would be overly simplistic.

Starting at X and going counter clockwise to O, then triangle etc is like the default usage% if you get my gist. X is almost always fastball because that is everyone's primary pitch not named Dickey. But individual pitch confidence will play a role as a game is played. If you sit on the O pitch and crush a couple of them, the confidence in that pitch will drop and it will alter the CPU behavior on when that pitch is thrown (and how often). As far as pitch edits are concerned, it is simply best to marry the usage with the correct button e.g. X first pitch O 2nd most used pitch and so on.

My understanding is that when physically playing the game against CPU, the break rating has a direct correlation to whiff %...so it is a combined visual effect of how much say a slider breaks but also a K/9 rating for that pitch as well. I don't know if you accounted for that or if Sean ever did?

[QuestGAV]

Re: Bastardized Pitch Edits 2015

Starting at X and going counter clockwise to O, then triangle etc is like the default usage% if you get my gist. X is almost always fastball because that is everyone's primary pitch not named Dickey. But individual pitch confidence will play a role as a game is played. If you sit on the O pitch and crush a couple of them, the confidence in that pitch will drop and it will alter the CPU behavior on when that pitch is thrown (and how often). As far as pitch edits are concerned, it is simply best to marry the usage with the correct button e.g. X first pitch O 2nd most used pitch and so on.

Yeah that's already accounted for.

My understanding is that when physically playing the game against CPU, the break rating has a direct correlation to whiff %...so it is a combined visual effect of how much say a slider breaks but also a K/9 rating for that pitch as well. I don't know if you accounted for that or if Sean ever did?

I didn't account for it and I don't think Sean did either. The way I'm doing guys now it would be relatively trivial to account for going forward but I'm not sure I'd want to even if that's the dev's intention. Among other things, I'm finding that it takes quite awhile for the noise to drain out of individual pitch outcomes. They're pretty reliable for guys who have thrown thousands of pitches but movement numbers are much more consistent in relatively small sample sizes.

[OhioCub]

Re: Bastardized Pitch Edits 2015

Starting at X and going counter clockwise to O, then triangle etc is like the default usage% if you get my gist. X is almost always fastball because that is everyone's primary pitch not named Dickey. But individual pitch confidence will play a role as a game is played. If you sit on the O pitch and crush a couple of them, the confidence in that pitch will drop and it will alter the CPU behavior on when that pitch is thrown (and how often). As far as pitch edits are concerned, it is simply best to marry the usage with the correct button e.g. X first pitch O 2nd most used pitch and so on.

My understanding is that when physically playing the game against CPU, the break rating has a direct correlation to whiff %...so it is a combined visual effect of how much say a slider breaks but also a K/9 rating for that pitch as well. I don't know if you accounted for that or if Sean ever did?

If what you're saying about break rating being tied to whiff percentage then scea did a terrible job of implementing that. I just looked up some of the pitches that had the highest whiff percentages last year and compared them to scea's ratings for those pitches. Some looked like it could be but others had break ratings in the 50's.