Stuff vs. Movement vs. Control

[aaronhelman]

Getting ready to do an experiment and wanted to get input to make sure I do everything right.

I'm going to set up a league with eight identical teams to play 162 times over several seasons.

I will clone exactly five pitchers with the following ratings:

S: 20 M: 20 C: 20
S: 20 M: 12 C: 12
S: 12 M: 20 C: 12
S: 12 M: 12 C: 20
S: 12 M: 12 C: 12

The goal is to see which (if any) of the three ratings will have a higher impact on overall performance. This information could provide significant value when drafting pitchers.

What else should I add to this experiment to add value?

[DCG12]

How are you accounting for number of pitches, pitch quality and pitch type.

There's also the fact that Markus has said previously that there are some hidden ratings such as I believe a Babip rating for all pitchers.

While the results might be interesting I'm not sure they will be definitive.

[jpeters1734]

Nice. I can't wait to see the results. I've have wondered what the dominate rating is. Another thought is what two pair best with each other. For example, i've heard that movement and control work best because it will improve stuff.

[RchW]

Originally Posted by jpeters1734

Nice. I can't wait to see the results. I've have wondered what the dominate rating is. Another thought is what two pair best with each other. For example, i've heard that movement and control work best because it will improve stuff.

If that were so then a pitcher with 10-20-20 ratings could not exist. They certainly do.

I suspect that the interaction of SMC is more complex internally. Remember that under the hood the scale is 250 with the last 50 being over the top displayed rating. That means a 20 rating can have a 25% variation if you choose not to display ratings above max. I'd certainly recommend that the OP not use 250 to generate the 20 ratings as these are very rare and would not represent the possible variation accurately.

[RchW]

Originally Posted by aaronhelman

Getting ready to do an experiment and wanted to get input to make sure I do everything right.

I'm going to set up a league with eight identical teams to play 162 times over several seasons.

I will clone exactly five pitchers with the following ratings:

S: 20 M: 20 C: 20
S: 20 M: 12 C: 12
S: 12 M: 20 C: 12
S: 12 M: 12 C: 20
S: 12 M: 12 C: 12

The goal is to see which (if any) of the three ratings will have a higher impact on overall performance. This information could provide significant value when drafting pitchers.

What else should I add to this experiment to add value?

I suggest you avoid using 250 to generate the 20 ratings as such players are rare. You might just add a super 20 (250) pitcher to the list to see the effect.

With no disrespect I doubt you will find a definitive answer. There is no way that the game could have maintained it's dominant position simming baseball if one set of ratings was definitive.

[aaronhelman]

Okay, the results are in, and they were shockingly more definitive than I expected. I'll put this together pretty briefly, but if anyone has specific questions, just let me know. Here's the simplest version of the results:

STUFF > CONTROL > MOVEMENT

I simmed eight seasons, reviewed stats after each season and then looked at career stats.

Every single one of the "stuff" guys ended up with a better career than either the "movement" or "control" guys (considering WHIP, FIP, and WAR). The affect of stuff was undeniable.

The control guys exhibited a less pronounced advantage over the movement guys, but it was still there. The control guys were only slightly better than the movement guys cumulatively.

However, on a season-by-season basis, the control guys were more consistent. The movement guys were more likely to just have an awful season here and there.

[jpeters1734]

How did you analyze the data? Also what constituted a better career?

[aaronhelman]

I looked specifically at WAR, WHIP, and ERA+ after each season.

I measured careers using the same criteria.

[jpeters1734]

Originally Posted by aaronhelman

I looked specifically at WAR, WHIP, and ERA+ after each season.

I measured careers using the same criteria.

Could you provide a spreadsheet with the career stats, and if possible, by pitcher type?

[Izz]

Were there any pitchers rated 15/15/15, or something similar?

[mpejkrm]

That's the order I figured it mattered, because it most models what pro teams look for in real life. Nowadays, it's hard for a draft prospect to get high on the board if he has below-average stuff. A lot of projectable HS guys go high because of great stuff, even in the control needs serious work. It's good to see that OOTP realizes that and inputs it into the game.

[fhomess]

Originally Posted by aaronhelman

The movement guys were more likely to just have an awful season here and there.

This isn't surprising since movement is most directly correlated with home runs allowed. Since pitchers will strike out upwards of 200 batters a year, walk maybe half that at the worst end of the spectrum, but give up only about 30 HR's, a particularly lucky or unlucky season with home runs will have a bigger percentage impact on their season.

I think the other thing to look at in detail is what the ratings are relative to league average. As you probably know, a league in which everyone has 16 stuff will result in a less dominant 20 stuff pitcher than a league in which everyone has only 10 stuff. The average stuff rating isn't always the same as the average control or average movement rating. In fact, they're often different by a couple of ratings points.

Overall, though, I think the conclusions you found make sense to me intuitively and based on my anecdotal experience.

[BusterKing]

Like Dcg12 said in the second post "How are you accounting for number of pitches, pitch quality and pitch type."

That would have an impact on the results also.

[hluraven]

What about defence? Was that equalised as it can affect many pitching categories

[redsox19934]

Originally Posted by aaronhelman

Okay, the results are in, and they were shockingly more definitive than I expected. I'll put this together pretty briefly, but if anyone has specific questions, just let me know. Here's the simplest version of the results:

STUFF > CONTROL > MOVEMENT

I simmed eight seasons, reviewed stats after each season and then looked at career stats.

Every single one of the "stuff" guys ended up with a better career than either the "movement" or "control" guys (considering WHIP, FIP, and WAR). The affect of stuff was undeniable.

The control guys exhibited a less pronounced advantage over the movement guys, but it was still there. The control guys were only slightly better than the movement guys cumulatively.

However, on a season-by-season basis, the control guys were more consistent. The movement guys were more likely to just have an awful season here and there.

currently in 2025 and pitcher struck out 325 batters with 19 stuff/12 movement/18 control.

[RchW]

Originally Posted by BusterKing

Like Dcg12 said in the second post "How are you accounting for number of pitches, pitch quality and pitch type."

That would have an impact on the results also.

Originally Posted by hluraven

What about defence? Was that equalised as it can affect many pitching categories

Not only that but the manual clearly states that stuff is an amalgam of pitch ratings and velocity. The pitchers created might not exist based on OOTP's documentation.

For example pitcher 2 given poor M and C ratings has unrealistically high stuff. Where does it come from? Average movement and control = average pitches hence limited contribution to stuff. You can't just create stuff out of average movement and control. He will obviously show stuff as the winner in any experiment given these ratings. Similarly pitchers 3 and 4 have unrealistically low stuff which will further confirm stuff as the key element in an experiment. High M or C ratings come from better pitches (quality, location, strikes etc) so this must add to the stuff rating as per the manual.

The original pitcher table.

S: 20 M: 20 C: 20 = 60 Good control pitcher
S: 20 M: 12 C: 12 = 44 This pitcher can't get 20 stuff from M12 and C12. See pitcher 1. If pitcher 2 exists pitcher 1 has to have S = 30 or more on a 20 scale.
S: 12 M: 20 C: 12 = 44 This pitcher has to get better stuff from M20 pitch ratings. How much is disputable.
S: 12 M: 12 C: 20 = 44 This pitcher has to get better stuff from C20 pitch ratings. This may not be equal to M and is disputable.
S: 12 M: 12 C: 12 = 36 Good control pitcher.

A revised pitcher table is suggested.

The 3 pitchers in the middle should have total ratings adding up to 48 not 44. This is the mid point and average of 60 and 36. When one uses 48 to distribute the ratings the pitchers become more realistic because they incorporate the contribution from movement and control. I realize this is arbitrary. Feel free to tear it apart.

S: 20 M: 20 C: 20 = 60
S: 16 M: 16 C: 16 = 48
S: 16 M: 20 C: 12 = 48
S: 16 M: 12 C: 20 = 48 maybe 14-14-20?
S: 12 M: 12 C: 12 = 36

I'm not saying I have this figured out, just suggesting that the OOTP definition of stuff means one has to account for it based on the other ratings. My choice of equalizing the contribution of C and M may be wrong for example. However please note that pitchers 2, 3, 4 can now be compared on the differences in M and C. If stuff controls pitcher 2 should show it.

Just my

[Sweed]

God bless anyone that wants to do a study like this. After all it's your game and you can do what you want with it. Though, I just can't for the life of me figure out why they would want to? To get an advantage in an online league or over the AI in a solo league? I'd like to think there is enough under the hood variance that even a massive study would not give a definitive answer (and no, I haven't read any of the results because frankly, I don't want to know). But if it does how much fun is your game once you have that information?

IRL you certainly can't do this study and Markus is trying to create the illusion of "real life" in OOTP. There is no real life scout, coach, GM, owner, or human being that can say "give me the guy with combination X,Y,Z stuff\movement\control because I've tested extensively and can say that is the best combo anyone can have."

The fun of OOTP too me is having the illusion that I am dealing with human beings and not a bag of ratings tied to some code. If one knows the inner workings the illusion is destroyed and the game just becomes a math problem.

This is just one guys opinion. I'm not trying to insult or disparage the OP as I said he is entitled to do whatever he wants with his game.

[louisvillec271]

Just last year a forum member conducted a ratings study which seems to be more structured and detailed than the one performed by the OP.

Here's a link to that thread.

[RchW]

Thanks for that. I knew that linearity and weight were big factors but couldn't remember that thread.

[Hajew86]

Originally Posted by aaronhelman

Okay, the results are in, and they were shockingly more definitive than I expected. I'll put this together pretty briefly, but if anyone has specific questions, just let me know. Here's the simplest version of the results:

STUFF > CONTROL > MOVEMENT

I simmed eight seasons, reviewed stats after each season and then looked at career stats.

Every single one of the "stuff" guys ended up with a better career than either the "movement" or "control" guys (considering WHIP, FIP, and WAR). The affect of stuff was undeniable.

The control guys exhibited a less pronounced advantage over the movement guys, but it was still there. The control guys were only slightly better than the movement guys cumulatively.

However, on a season-by-season basis, the control guys were more consistent. The movement guys were more likely to just have an awful season here and there.

Great stuff. Interestingly enough I've been going after guys with high control ratings for quite some time and they've always seemed to work well for me. I can definitely understand how stuff would trump all though.