Player Speed Ratings

[Dyzalot]

Originally Posted by Reed

You are misinterpreting what Matt said. If a player is generated with excellent outfield range but no speed means he is fast or can quickly judge where a ball is going to etc., but for some reason he will not attempt many stolen basis. Maybe he can not get a good read on pitchers or he can not slide avoiding tags, etc..
Just because a player has good defensive range does not mean they will attempt a lot of stolen bases.

I misinterpreted nothing. Matt says that speed rating also affects infield hits, advancing bases and stealing yet they derive the number solely from stealing results. That seems like a huge imbalance.

[t-bone shuffle]

It seems to me the root issue here is the game's need to produce statistical accuracy in replay mode while balancing player modeling/ability/potential in the fictional mode (which includes current season MLB rosters which become "fictional" when advancing one day).

In a replay, there's likely always the player who had a 13/1 SB/CS season, but is average or below average in measured footspeed. How do you reconcile this reasonably, particularly in a fictional setting?

The only solution I can imagine is to add a 4th rating that is strictly devoted to player footspeed. This would address the issues like a slow dribbler to the left side, or stretching a single into a double. With this type of rating, the afore mentioned Willie Mays (in his prime) would carry a rating that causes his success/safe rate to be well above league average. Boog Powell's success rate would be in the zero neighborhood. (Yes I saw them both in their prime).

Then consider that this footspeed rating would somehow need to combine/work with a "baserunning" ability to affect a "good" baserunner from a "bad" one on the various plays where both come into play.

So, a lot to ask/expect/hope for. Unfortunately not likely. Not necessarily a bad thing though, IMO.

[Anyone]

They need separate ratings for running speed and the likelihood of a stolen base attempt. Simple as that.

For historical players, stats for the pure speed rating were available before more advanced stats existed, Bill James writing in the 1980's had "speed scores." I'm not sure exactly what was in them, but something like stolen bases, stolen base percentage, triples as a percentage of hits (or relative to doubles)-- which by default speed is used for already, % of AB's (or of outs) that result in grounding into a DP, Runs scores as a percentage of times on base not counting HR's.... Also, how far back does percentage of times taking an extra base go?

Any of those alone isn't a perfect indicator or as good as what we have now. Each has distortions. But put them together and the distortions will tend to cancel each other out.and give a better approximation of a players' raw speed than what we have in the game now.

[Argonaut]

Originally Posted by Anyone

They need separate ratings for running speed and the likelihood of a stolen base attempt. Simple as that.

For historical players, stats for the pure speed rating were available before more advanced stats existed, Bill James writing in the 1980's had "speed scores." I'm not sure exactly what was in them, but something like stolen bases, stolen base percentage, triples as a percentage of hits (or relative to doubles)-- which by default speed is used for already, % of AB's (or of outs) that result in grounding into a DP, Runs scores as a percentage of times on base not counting HR's.... Also, how far back does percentage of times taking an extra base go?

Any of those alone isn't a perfect indicator or as good as what we have now. Each has distortions. But put them together and the distortions will tend to cancel each other out.and give a better approximation of a players' raw speed than what we have in the game now.

That's a really good point, and something I didn't think about. Surely with all of those things together, a proper Speed rating could be sussed out. I'm totally on board.

[One Great Matrix]

Originally Posted by Pirates

The problem here is this:
I have seen many fast baseball players who did not steal a lot of bases, have an abysmal running speed.

This is factually incorrect.

There's a tton oof players in the game & they don't know everyone of them from your seat. When notified of an error they look into it.
From what I heard Mantle was a fast runner around the bases, his stealing rating is rather high on one of my PT cards but his speed is VERY low.

I think the basic batting & pitching statistics are least error-prone.

[One Great Matrix]

Originally Posted by t-bone shuffle

It seems to me the root issue here is the game's need to produce statistical accuracy in replay mode while balancing player modeling/ability/potential in the fictional mode (which includes current season MLB rosters which become "fictional" when advancing one day).

In a replay, there's likely always the player who had a 13/1 SB/CS season, but is average or below average in measured footspeed. How do you reconcile this reasonably, particularly in a fictional setting?

The only solution I can imagine is to add a 4th rating that is strictly devoted to player footspeed. This would address the issues like a slow dribbler to the left side, or stretching a single into a double. With this type of rating, the afore mentioned Willie Mays (in his prime) would carry a rating that causes his success/safe rate to be well above league average. Boog Powell's success rate would be in the zero neighborhood. (Yes I saw them both in their prime).

Then consider that this footspeed rating would somehow need to combine/work with a "baserunning" ability to affect a "good" baserunner from a "bad" one on the various plays where both come into play.

So, a lot to ask/expect/hope for. Unfortunately not likely. Not necessarily a bad thing though, IMO.

Yes whenever I've noticed something off I've also thought of sort of contradictions or just plain limitations in coding. Re-creating baseball isn't cake.

But Bill James best work is great stuff. Some of it...I do not care for but his best stuff could surely assist in little details.

[Drstrangelove]

I'm wondering if this is simply a request to change what the game says as opposed to what it does?

I.e., so far, no one has said that Clemente's or Mantle's number of SB or triples are wrong? The complaint seems to be that the game should say "FAST" where it doesn't? Is that the complaint? I.e., if the game had said "FAST" next to those players, we wouldn't have a thread?

[Reed]

Originally Posted by Drstrangelove

I'm wondering if this is simply a request to change what the game says as opposed to what it does?

I.e., so far, no one has said that Clemente's or Mantle's number of SB or triples are wrong? The complaint seems to be that the game should say "FAST" where it doesn't? Is that the complaint? I.e., if the game had said "FAST" next to those players, we wouldn't have a thread?

Basically the same thing I was wondering. Are Clemente/Mantle OOTP total doubles and triples close to their real life numbers or is there a huge discrepancy, assuming the same number of at bats. People talk about the various ratings but no one has posted the actual stat output.

[Dyzalot]

Originally Posted by Drstrangelove

I'm wondering if this is simply a request to change what the game says as opposed to what it does?

I.e., so far, no one has said that Clemente's or Mantle's number of SB or triples are wrong? The complaint seems to be that the game should say "FAST" where it doesn't? Is that the complaint? I.e., if the game had said "FAST" next to those players, we wouldn't have a thread?

Matt said that the speed rating is used for things like beating out infield hits, whether you gidp or not, stretching doubles into triples, etc.

[Drstrangelove]

Originally Posted by Dyzalot

Matt said that the speed rating is used for things like beating out infield hits, whether you gidp or not, stretching doubles into triples, etc.

It just seems some suggestion in-game stats are wrong or must be wrong. I was wondering if that's conjecture or something they can post. It's simply hard to move an issue without evidence.

[Dyzalot]

Originally Posted by Drstrangelove

Is anyone complaining that anyone's actual infield hit counts, GIDP counts are wrong either? The question is: Is the only thing that people are saying is that the WAY a player's speed is being described is wrong?


Not the player's triples, stolen bases, infield hits, stolen base percentage, GIDP, putouts (in the field), etc. None of the stats. Is the complaint that it says one thing in a box on a screen, and in spite of the stats matching real life stats, it doesn't say the thing in the way a person thinks it should say it?

Well if the results are affected then either the speed rating is being displayed incorrectly or there is a bug affecting the way speed is supposed to influence those things.

EDIT: After thinking about it a bit more, I suppose one other possibility is that "speed" in OOTP is a rating that actually encapsulates several different kinds of speed ratings that are "under the hood" and if the overall "speed" rating is heavily weighted by the sub component of the "stealing speed" rating then one could presumably have a low rating when it comes to speed overall even though a few "subratings" are fairly high such as the ability to leg out an infield hit because the heavily weighted component of the overall "speed" rating is low. This could result in someone like Mays having a low overall speed rating due to a low stealing speed rating but still having high ratings for legging out triples, infield hits, etc. It would seem strange to me if the speed rating is broken down to that level in the game such that you couldn't use the overall rating as a gauge for how well a player would perform in any of the activities governed by that rating. But its possible. I hope that makes sense.

[TomVeal]

Perhaps "Speed" should be relabeled "Jump". It appears to me that other ratings, such as IF and OF range, reflect raw speed.

[Mannyw]

I'm the original poster. I've followed this thread and given it a lot of thouoght. I looked up Sabermetrics in relations and found something very interesting called Speed Score. Below is the information copied from Wikipedia:


https://en.wikipedia.org/wiki/Speed_Score


So I do think it is possible to more accurately calculate player speed. I played APBA Baseball growing up and have messed with their Window version some too. They have somehow calculated player speed for years, seemingly very accurately. In their system a player is ranked 1-20, with a slow player being 1-6, average 7-14 and fast 15-20. Only the most elite runners reach 20, Rickey Henderson, Tim Raines come to mind.

Their speed ranking is entirely separate from a runners SB success. In speed a player speed is subtracted from outfielder's arm rating and then that sum is subtracted from a number representing ball depth in outfield.

In base stealing the runner has a 0-36 base stealing rating. Catcher have a -4 to +6 arm rating and pitchers have a 0 to +3 pitch-off move rating. So a +4 catcher and a +1 pitcher would have a sum of +5. This +5 is subtracted from the base stealing number, say 30 of a possible 36. So 30 minus 5 gives the runner 25 out of 36 chances to successfully steal the base.

Just so thoughts, it would take work but runner speed accuracy is possible, even for non modern era players.

[Reed]

Originally Posted by Mannyw

I'm the original poster.


So I do think it is possible to more accurately calculate player speed. I played APBA Baseball growing up and have messed with their Window version some too. They have somehow calculated player speed for years, seemingly very accurately. In their system a player is ranked 1-20, with a slow player being 1-6, average 7-14 and fast 15-20. Only the most elite runners reach 20, Rickey Henderson, Tim Raines come to mind.

Their speed ranking is entirely separate from a runners SB success. In speed a player speed is subtracted from outfielder's arm rating and then that sum is subtracted from a number representing ball depth in outfield.

In base stealing the runner has a 0-36 base stealing rating. Catcher have a -4 to +6 arm rating and pitchers have a 0 to +3 pitch-off move rating. So a +4 catcher and a +1 pitcher would have a sum of +5. This +5 is subtracted from the base stealing number, say 30 of a possible 36. So 30 minus 5 gives the runner 25 out of 36 chances to successfully steal the base.

Just so thoughts, it would take work but runner speed accuracy is possible, even for non modern era players.

Okay let us get to the bottom line which is stat results. You may be right but to convince me you will need to show me where the stat output is different than their real life numbers which NO ONE has shown in the 3 pages on this thread. Actually the printout on the 1st page of this thread showing doubles, triples, attempted stolen bases, and stolen base seems reasonable to me compared to their real life stats. I do not see their number of grounded into double plays and that would be interesting compared to their real life numbers.

Again, show me where the stat outputs are wrong, not isolated examples but a large sample size.

[Argonaut]

Been poking around a bit, and found datapoints from fangraphs and almanac:

https://tht.fangraphs.com/predicting-double-play-rate/
https://www.baseball-almanac.com/pla...php?p=mayswi01

According to this, Willie Mays' expected number of GIDP in 1962 was 32, but in actuality he hit for 19 GIDP. So the difference does suggest his speed helped him to avoid more double plays.

I'm at work now and can't run a simulation, but it would be interesting to see what OOTP would say Willie's GIDP output would be.

This is important to get right, because as I've seen elsewhere, GIDP destroys value:

https://www.beyondtheboxscore.com/20...duction-killer

This article shows Brooks Robinson going from above average at the plate to below average, just from how often he GIDPed.

[Mannyw]

I'm primarily talking about player speed as a basis of ability to take an extra base, score from first on a double or score from second on a base hit.

Doubles, triples and stolen bases is a separate stat, we all clearly know that.

Reed, you ask for a comprehensive analysis to listen to this concern. Well, the comprehensive data is on Statcast, compare player speed ratings in OOTP to the below data and you have your data:

https://baseballsavant.mlb.com/sprin...=&team=&min=10

As a manner of analysis to determine non-Statcast era player speed read the formula thread I listed, you like data, there is some data for you....but actually read and consider the points being made here:

https://en.wikipedia.org/wiki/Speed_Score

I love this game and I'm suggesting ways to make it better, not attacking it. Player speed is a weakness in OOTP Baseball and I'm suggesting ways that might be considered to help improve this rating.

[Anyone]

Yes, taking an extra base (including tagging up, 2nd to 3rd on a ground ball, or the front end of an intended sac bunt) and turning ground ball DP's into fielders' choices are the keys to having a real speed rating that we don't have, that Willie Mays' speed rating being lower because managers didn't send him to try to steal bases-- as much as anything because in his fastest years few players attempted steals-- doesn't allow him to do the way he could.

Just giving him a good basrunning rating so he makes good decisions isn't enough there. I've read bits and pieces about Mays that suggest he did in fact make good baserunning decisions. But his great speed, especially when young, makes the "right" decision being going for the extra base a lot more often. That's what needs to be simulated.

I'd also throw in infield hits, though outside of bunting for a hit those are more flavor and not quite as important, because the engine will make sure his batting average will work out right-- but flavor matters also.

I rarely play historical, though usually import a few historical players into fictional leagues, but I'm less affected by this. But it just makes no sense. If as a GM you wanted your real life team to attempt more steals, maybe getting faster players would be on your list, but not at the top. You'd look for a manager who likes to give the green light and players who were talented at base stealing, including talented at getting a good jump. Usually base stealing talent would correlate with speed, though it's not the same thing, and getting a good jump and the willingness to take risks are very different.

For historical leagues to deal with players who tend to get 15 steals a year and are only caught once, it should use a 3 year period or something-- half of those are probably one-year flukes-- but those who did that consistently probably weren't great base stealers who only stole when they thought they caught the pitcher ignoring them, though that was its own talent, to notice when they had a good opportunity. You might give a low "jump" rating and a great stealing rating, which I've seen other games do, though maybe ideal realism would be a mediocre stealing rating but a high "opportunism" rating, to notice when the pitcher wasn't properly holding them (or they weren't being held on at all).

Speed, while overrated before sabermetrics became more mainstream, had a lot of uses, its main uses, that didn't involve stolen bases.

[Drstrangelove]

Originally Posted by Anyone

Just giving him a good basrunning rating so he makes good decisions isn't enough there. I've read bits and pieces about Mays that suggest he did in fact make good baserunning decisions. But his great speed, especially when young, makes the "right" decision being going for the extra base a lot more often. That's what needs to be simulated.

The game allows managers to tell individual runners (under strategy) to run all the time or be as aggressive as possible on the bases. So, really, this is a choice that every manager can already make.

[Reed]

So instead of talking about speed rating we are actually talking about the baserunning rating.
Manny, I did read the statcast site you referenced above and it said the primary, two thirds, part of the formula is “runs”. Do you know if OOTP uses runs to determine their “baserunning” rating. If not, what do they use to determine their baserunning rating?
How do they determine between Sam Crawford and Buck Freeman who has the better baserunning rating?

[kingusc]

Can somebody try to bottom line this for me?

a) If I'm scouting a draft and want the best team speed, I should look first to the _______ rating?

b) If I want to steal a base, the __________ rating is most important? (I believe this to be STE, with SPD factoring in as well).

c) If I want to know who is the fastest on my team from home to 1B, I should look to the _________ rating?

d) If I want to know what players may make the worst mistake on the basepaths, I should look primarily to the _________ rating? (I believe this to be Baserunning, later to be named the Puig rating).