Pitching - Stuff & Cont, Cont & Move, Move & Stuff

[Long_Long_Name]

Anyone going to dig up the study? I can't seem to find it

[james17]

Stuff then Movement then Control.

[Carlton_Willey]

Originally Posted by Elendil

I've found that Movement is the most important rating for preventing runs (which is what we're all interested in, right?). And that's a bit sad, b/c like others here, I think it's the only rating that doesn't correspond to anything IRL.

IRL a high movement, no control, no stuff pitcher is throwing a knuckle ball.

[obaslg]

Originally Posted by Joshv02

Interesting -- is that overall at all ratings, or only when ratings for each are > 5.

I found, fiddling with Jayzone's utility, that if ratings were =<5, than it is stuff/cont/mov, but that the curve appeared steep.

I'd imagine that a 3/3/3 (s/c/m) pitcher going to a 3/4/3 is better than a 3/3/4, but this is not true when we start at 6/6/6.

That's interesting. So...I assume you determined this with trials using his utility. That would mean that his utility has that factored in somehow, which is interesting.

That may answer the first question I asked him in that other thread: It may be better to do the projections in two steps (step one to determine basic stats like Ks and BBs and step two to determine ERA) than to go straight to ERA, because whatever formula he uses to project ERA from basic stats may have a curve in it - e.g., homers have an exponentially increasing effect.

I've never been very happy with my homemade pitcher projections, so I'm interested in other models. My model seems worst when the ratings are far out of line with each other, e.g. with 100+ Stuff and <50 Con and Move. How good a prediction do you think the utility gives you?

[obaslg]

Originally Posted by Long_Long_Name

Wasn,t there a study a while ago that found out stuff was clearly the most important of the three ratings? I seem to remember movement was slightly more important than control, but that stuff was miles ahead.

There was, though it wasn't done with regression. From what I've seen, all regressions of the basic ratings on final ERA have shown Stuff to be most important, but there may be other issues, as mentioned above.

[Long_Long_Name]

Originally Posted by obaslg

There was, though it wasn't done with regression.

I think it was...

[congobeast]

Control uber alles.

Control cuts down on your pitch count (fewer balls, thus fewer pitches and BB), and makes one pitcher give you more bang for your buck.

Key to this is at least a decent infield defense behind him, turning grounders into outs. With a poor defense, your pitch counts will shoot right back up again.

Of course, YMMV.

[Cooker]

This is a very interesting thread ....

I always thought it should be stuff, movement, then control. I hate being beaten by the proverbial 3-run HR (a la Earl Weaver). But I never considered the effect control has in keeping pitch counts down. Good point.

A SP in one of my solo leagues has 95 stuff, 30 control, and 90 movement. He has 94 endurance, and needs it. He's always over 80 pitches by the 6th inning. His talent for control is 70, so hopefully his control rating will catch up to his control talent someday. But right now, he's at best a .500 pitcher despite his high ratings

[Long_Long_Name]

I must bring up the study again. I can't find it .

[Johnny Slick]

Originally Posted by Carlton_Willey

IRL a high movement, no control, no stuff pitcher is throwing a knuckle ball.

I disagree. In my mind, stuff is the amount of actual movement in a pitcher's repertoire whereas movement is how well he moves pitches within the strike zone. See what I mean? A guy with a fastball with huge hops and a big breaking curve might have great stuff, even good control if he can keep the pitches within the strike zone, but if the fastball is always up and in and the curve is always low and away, he'll have a poor movement rating.

A knuckleball pitcher, then, might have good stuff, no velocity (which rightly has little impact on the game), and variable control and movement.

To bolster this argument in a way, check out what people had to say about the deadball era pitcher Grover Cleveland "Pete" Alexander. IIRC he was known as "Old Low and Away" because, well, every pitch he threw ended up low and away. In a homerun-dominant era, he'd get tagged for homeruns because people would know the general vicinity of all his pitches, and, if they got the timing and the optical illusions right, they could hit the ball a long way. Since Pete pitched mostly in a period where nobody hit homeruns anyway, he could get away with that sort of thing (and, IIRC, he moved pitches around in the strike zone late in his career when baseball did begin to get homer happy).

As far as what's important... this is a cagey answer, but it depends on what your league values. Stuff is the obvious #1 answer, but IMO it's not a lot more important than the other values. At least not as much as some leagues think it is. If everybody is going after pitchers with lots of stuff and you're the guy who takes players with good movement or control, you will, 9 times out of 10, have a fearsome pitching staff. For that reason, I often find myself favoring the latter 2 abilities, usually movement above all else because that seems to be the most underrated of the 3.

[Long_Long_Name]

I think that "movement" and "stuff" are quite nebulous. Isn't movement, like, the incapacity to have control? And stuff... is just... nothing.

[obaslg]

Originally Posted by Long_Long_Name

I must bring up the study again. I can't find it .

There's one that I started pretty recently that did deal with regression a lot - I thought you were talking about an earlier one. Search by my name and you'll find the recent one.

[Long_Long_Name]

Originally Posted by obaslg

There's one that I started pretty recently that did deal with regression a lot - I thought you were talking about an earlier one. Search by my name and you'll find the recent one.

The earlier one came out just a week or two after OOTP6 came out, I think.

As for yours, I'll take a look .

[Joshv02]

Originally Posted by obaslg

How good a prediction do you think the utility gives you?

I haven't tested it, but from a general feel, I find it to be accurate especially with pitchers.

[thomamon]

It seems that the best pitchers in the APBL seem to be those that have a high Stuff and Movement rating and a decent control rating.

Andy Strope who now has a 10 in stuff and movement and 6 in control led the AL in ERA. He also has a 9 in endurance.

TM

[Jayzone]

Originally Posted by Joshv02

I think this is the most recent version.

It converts ratings to projected stats, considering league totals. One of my favorite little tools.

First of all, thanks again for plugging and the kind words, Josh. And, yes, v0.93b is the most recent versions. Well, I have been busy recently (and will continue to be so). So, v0.93b is likely to be the version for the foreseeable future. Now back to the topic.

Originally Posted by obaslg

Thanks. I just asked him a couple of questions in that thread, since my regressions say that it's Stuff-Cont-Move.

Questions are answered in that very thread. The answer to which rating is more important really depends on the league enviroment. Low strikeout environment means stuff is less important. Low walk environment means control is less important. Low HR environment means movmement is less important. Under defualt setting, stuff and movment (if above replacement) are important. Altough, control should not be completely ignored and there is also an added bonus of being able to go deeper into the game (as mentioned earlier in this thread). Being purely political, I would say all raings are important.

Originally Posted by Joshv02

Interesting -- is that overall at all ratings, or only when ratings for each are > 5.

I found, fiddling with Jayzone's utility, that if ratings were =<5, than it is stuff/cont/mov, but that the curve appeared steep.

I'd imagine that a 3/3/3 (s/c/m) pitcher going to a 3/4/3 is better than a 3/3/4, but this is not true when we start at 6/6/6.

Under defualt setting, the rating to stats relationship for movement does indeed exhibit curvelinear behaviour. (I think I mentioned this in my original rating translator thread and/or in the documentation.) This means that a gain from 80 to 90 is more than 30 to 40. I don't know the reason for this but I designed the rating translator to be flexible/dynamic nevertheless. So, the translation for any rating would auto switch to curvelinear model if such model is tested to be better-fit. This means that if any other rating relationship exhibits curve-linear behaviour then the program will try to adjust to it. Another thing to note is that even other ratings seem to be in linear relationship, the extreme values (extreme low and high) can be not consistent with the linear relationship. This is most likely due to small sample issue. Still, this creates issue in determing where the true relationship is supposed to be at the extreme values.

Originally Posted by Joshv02

I haven't tested it, but from a general feel, I find it to be accurate especially with pitchers.

I haven't have a lot of feedback on the accuracy of Rating Translator and I would love to have more feedback on this aspect. As I stated in the documentation, my feel and limited testings tells me that the accuracy is reasonably satisfactory. Bearing in mind that 100% or close to 100% translation accuracy is impossible simply because the resulting actual stats are the results resulting from random events and there are also a lot of variables involved. It is not a simply matter of knowing y with y = a*x if a and x are known.

[sebastian0622]

Originally Posted by Jayzone

I haven't have a lot of feedback on the accuracy of Rating Translator and I would love to have more feedback on this aspect.

I would have been able to help you there, but some German commish stopped simming the league that I was primarily using it in!

To anyone who hasn't used it: this tool is especially handy if you're in a league that adjusts league totals or does something unique. Not that it bothers me if folks don't use it--the more of an advantage for me

[toolboxnj]

I'll go on a limb and say movement, stuff then control. A pitcher with 100 stuff and 100 movement is tough to beat.. have one on my solo leauge and has the lowest OPS in the NL. Just wish he had more endurance.

[DerekJeter2]

Movement is easily first, then Stuff 2nd. I've seen 10 stuff, 2 control, 10 movement pitchers have their ERA around 3 a lot. I haven't seen any 10 stuff, 10 control, 2 movement around a 3 ERA, they usually tend to have about a 4.50 ERA because they give up 50 HRs a year. Their is no question Movement is the most important, but with my expieriences Stuff comes in quite handy and control dosen't make that much of a difference.

[Jayzone]

Originally Posted by sebastian0622

I would have been able to help you there, but some German commish stopped simming the league that I was primarily using it in!

To anyone who hasn't used it: this tool is especially handy if you're in a league that adjusts league totals or does something unique. Not that it bothers me if folks don't use it--the more of an advantage for me

I would still love to hear what you (or anyone) think on the matter of accuracy issue if you ever have the chance to complie your thoughts.

And, oh, thanks for plugging as well. Well, following your logic, I probably should've never released the utility and kept it to myself if I were in some online leagues. Just kidding.