The Turn DP rating—does it do anything?

[ThePretender]

If a 2B has a DP rating of 1 they should be getting half or less than the proper amount of dps. That's essentially the equivalent of a 1B trying to start the DP. They should be noticeably screwing these plays up, either by making an error or only getting 1 runner.

[jpeters1734]

Originally Posted by SirSaab

Which 2b won the gold glove?

I'm also curious who won the GG

[Furious]

Originally Posted by ThePretender

If a 2B has a DP rating of 1 they should be getting half or less than the proper amount of dps. That's essentially the equivalent of a 1B trying to start the DP. They should be noticeably screwing these plays up, either by making an error or only getting 1 runner.

I don’t know if half is the right percentage of plays they should be screwing up, but I agree it should be a very large, very noticeable percentage.

[ThePretender]

Originally Posted by Furious

I don’t know if half is the right percentage of plays they should be screwing up, but I agree it should be a very large, very noticeable percentage.

You're right more than half.

[Garlon]

With regards to the reference of Bill Mazeroski, he turned 1706 DP in his career, and according to the formulas from Win Shares this was about 142 more Double Plays than the average 2B during the seasons he played. This is 9% more than average. Part of turning a DP is having a good DP partner too. If you figure "Turn DP" means that you are the player making the second assist, then roughly half the time the 2B will be the "turn" partner. So if you pair Mazeroski with an average partner, that makes his totals roughly 4.5% better than average. In the example in the original post it seems that this is approximately what was happening. Perhaps this test could be done again with both the 2B and SS Turn DP ratings increased instead of just the 2B. It seems that the results seem fine and that nothing should be changed with the game without further evidence suggesting something is amiss. Remember that in the original test you are actually putting a legitimate 2B at the position, not a player who is out of position. So the relative differences between the 2B in the test seem fine to me.

Even if you do not turn the double play you will still get 1 out. The second out adds about another 0.32 runs saved from the information I have read on the subject.

As for Mazeroski, he had about 335 more assists in his career than the average 2B, so maybe 70 of those were on the double play and the rest were on other plays.

With regards to what Leo the Lip was saying about Philadelphia and Joost, those teams from 1949-1951 were 19%, 5%, and 8% better than league average in turning double plays when you account for double play opportunities. Joost was +64 double plays turned for his career, which was about 7.5% better than average.

Regarding Nap Lajoie, he was about +41 double plays turned for his career, and he turned 1050 double plays. His defensive statistics have to be interpreted in the context that he played for poor defensive teams. He made about 109 more assists than average in his career, but he made 178 fewer errors than average. So when you consider those two things you realize that the +109 assists are really the result of preventing errors since his assists above average are less than his errors saved.

[ThePretender]

But that begs the question of what 1 DP should be. Should it be 1 DP compared to 2B so that it's just on the lower end of what's acceptable, or should 1 DP skill basically mean you're terrible at it and should be forced off the position because you can't do it? To me it should be the second one, because it refers to all players. We aren't just comparing turn dp to all 2B, we are looking at the rating for all players, and 1 should mean the player simply cannot complete a dp.

[Rain King]

Does anyone know of a historical example where a player was a solid 2nd baseman except for the ability to turn a double-play?

I'm not convinced it is all that difficult of a skill when you already possess the ability to move laterally and field a ground ball at a professional level.

It seems like one of those skills that is easy to become competent at but very difficult to become elite at and it seems like the game is representing that pretty well.

[RchW]

Originally Posted by OutS|der

When looking at numbers like this can we use the players actual ratings under commissioner rather then going off scouted ratings which could be completely off.

A 1 out of 5 means nothing to me but the actual number does.

Also totals are affected by league totals so different totals will result in different stats even with the same ratings.

IE give a player 250 rating in power and 2019 totals and he’ll hit a lot more HR’s then if you used the 1994 totals. So just saying the rating is useless leaves out a lot of factors

Two very critical points.

[Leo_The_Lip]

Originally Posted by Garlon

With regards to the reference of Bill Mazeroski, he turned 1706 DP in his career, and according to the formulas from Win Shares this was about 142 more Double Plays than the average 2B during the seasons he played. This is 9% more than average. Part of turning a DP is having a good DP partner too. If you figure "Turn DP" means that you are the player making the second assist, then roughly half the time the 2B will be the "turn" partner. So if you pair Mazeroski with an average partner, that makes his totals roughly 4.5% better than average. In the example in the original post it seems that this is approximately what was happening. Perhaps this test could be done again with both the 2B and SS Turn DP ratings increased instead of just the 2B. It seems that the results seem fine and that nothing should be changed with the game without further evidence suggesting something is amiss. Remember that in the original test you are actually putting a legitimate 2B at the position, not a player who is out of position. So the relative differences between the 2B in the test seem fine to me.

Even if you do not turn the double play you will still get 1 out. The second out adds about another 0.32 runs saved from the information I have read on the subject.

As for Mazeroski, he had about 335 more assists in his career than the average 2B, so maybe 70 of those were on the double play and the rest were on other plays.

With regards to what Leo the Lip was saying about Philadelphia and Joost, those teams from 1949-1951 were 19%, 5%, and 8% better than league average in turning double plays when you account for double play opportunities. Joost was +64 double plays turned for his career, which was about 7.5% better than average.

Regarding Nap Lajoie, he was about +41 double plays turned for his career, and he turned 1050 double plays. His defensive statistics have to be interpreted in the context that he played for poor defensive teams. He made about 109 more assists than average in his career, but he made 178 fewer errors than average. So when you consider those two things you realize that the +109 assists are really the result of preventing errors since his assists above average are less than his errors saved.

Thanks for this! I've been looking for data such as this without any luck. May I ask where you found it?

I'm also looking for quantitative data on the opportunities to turn groundball DPs. Basically, how often does a batter come up with a man on first (at least) and less than two outs? Because in all other situations the 'Turn DP' rating doesn't matter at all.

[Isryion]

Originally Posted by Garlon

. Remember that in the original test you are actually putting a legitimate 2B at the position, not a player who is out of position. So the relative differences between the 2B in the test seem fine to me.
.

The fact that they were labeled as 2B is irrelevant to their DP rating in that fielding ratings are not at all relative to players at the same position. They are overall/absolute ratings of a players skill at that ability. Position ratings are a composite of many things and include experience, but no amount of experience is going to make a player with a 1 arm a competent 3B.

The original player in question turned DPs just fine immediately after being moved to second base. I'd bet that you might be able to take a first baseman, pitcher, or LF, who has a 1 DP rating and he'd be fine at turning double plays at 2B (though I don't know if throwing errors would come into play there and he'd otherwise be terrible).

As The Rain King and The Pretender point out, if the skill isn't that difficult or relevant, maybe it could be based off of error? You're unlikely to play a low error player in the middle infield. As I mentioned above, a 1 in a skill usually means a player doesn't really have that skill/ability at all in OOTP.

[Garlon]

It matters if a player is a 2B or not because they also have an "experience" rating in the editor. Try putting a player there with 0 experience at 2B for a season with that minimum DP rating and see what happens.

Rain King, I think Sandberg is an example as the Cubs were about 10% below average in turning DP during his seasons, but remember it takes a good DP partner so it is not all on Sandberg.

Leo The Lip, I used the Bill James formulas from Win Shares to compile the data for every team and player since 1871. Basically the formulas look at things like estimated players reaching 1B against the team, and the team assist totals as an approximation of groundball percentage, and then compare the league rate to the team rate to determine how many double plays they were expected to make in a season.

[Furious]

Originally Posted by Garlon

With regards to the reference of Bill Mazeroski... It seems that the results seem fine and that nothing should be changed with the game without further evidence suggesting something is amiss.

The results aren’t fine even if you only focus on the Bill Mazeroski end of the spectrum and ignore what the study was actually trying to test. The “Bill Mazeroski” in the study (the guy with a 200 Turn DP rating), turned an average of six double play a year more than the league average. That’s not 9%.

But forget about Bill Mazeroski; I wasn’t trying to test whether the game accurately simulates the skills of a Bill Mazeroski. The point of the study was to see what the game did with an “anti-Mazeroski”, a player with extremely poor double play skills. The tests indicated that such a player is only 10% worse than someone who is better than average (remember, the anti-Mazeroski had a Turn DP rating of 30 and the four other non-Mazeroski 2Bs had Turn DP ratings of 140). That doesn’t seem “fine” to me at all; that seems very much like something is amiss. 30 is not 10% less than 140.

[Furious]

Originally Posted by Garlon

It matters if a player is a 2B or not because they also have an "experience" rating in the editor. Try putting a player there with 0 experience at 2B for a season with that minimum DP rating and see what happens.

I believe that is exactly what Isryion did. He had a guy with 4/4/4/1 ratings (1-5 ratings scale) and he put him at second. Isryion can correct me if I’m in error about this, but I believe the guy had no 2B rating at all until he put him there. After a year or two playing second in the minors, he brought him up as a 4-rated 2B; still with a 1 Turn DP rating.

[Furious]

Originally Posted by OutS|der

A 1 out of 5 means nothing to me but the actual number does.

Also totals are affected by league totals so different totals will result in different stats even with the same ratings.

Originally Posted by RchW

Two very critical points.

The first very critical point was addressed in the original post. The 1-rated Turn DP 2B’s actual number was 30, the four 4-rated Turn DP 2Bs’ actual numbers were 140, and the 5-rated Turn DP 2B’s actual number was 200.

As for the second very critical point, this was a ten-season study with EVERYTHING exactly the same for every team in every season EXCEPT the Turn DP rating of the second baseman. The League Totals Modifiers never changed.

[Furious]

Originally Posted by ThePretender

But that begs the question of what 1 DP should be. Should it be 1 DP compared to 2B so that it's just on the lower end of what's acceptable, or should 1 DP skill basically mean you're terrible at it and should be forced off the position because you can't do it? To me it should be the second one, because it refers to all players. We aren't just comparing turn dp to all 2B, we are looking at the rating for all players, and 1 should mean the player simply cannot complete a dp.

That’s the way I interpret the component ratings. If you can’t do something at all you get a zero rating; non-catchers have zero ratings at catcher skills. If you can do something a little, but far below average, you get a very low rating; perhaps 20-30 on a 20-80 scale, 1-4 on a 1-20 scale, 1 on a 1-5 scale. If you’re far below average at turning a double play, you shouldn’t be playing second base, but if your manager insists on putting you there, your double play statistics should be noticeably bad.

[Ktulu]

Originally Posted by ThePretender

But that begs the question of what 1 DP should be. Should it be 1 DP compared to 2B so that it's just on the lower end of what's acceptable, or should 1 DP skill basically mean you're terrible at it and should be forced off the position because you can't do it? To me it should be the second one, because it refers to all players. We aren't just comparing turn dp to all 2B, we are looking at the rating for all players, and 1 should mean the player simply cannot complete a dp.

As mentioned the game accounts for this, you can play around in the editor and adjust the component ratings for all positions and see how they affect a players ability to play a position. There's definitely an argument that some of the thresholds should be adjusted.

If you’re far below average at turning a double play, you shouldn’t be playing second base, but if your manager insists on putting you there, your double play statistics should be noticeably bad.

If you have the skills to play the position at all you have a baseline ability to turn double plays. The rating is there to help distribute DPs appropriately between the skill levels of major league players. It's not a representation of a player's physical ability to make the pivot at second base. You can easily see this by looking at the correlation of DPs by 3B and their turn DP rating.

If you have some real life data to show that the range of DPs is wrong then we could determine if something is wrong. A 10 season sample of one player type (that the game may not even create on it's own) is a start, but people like Garlon have edited thousands of historical players to produce accurate results and others have simmed hundreds of seasons to make sure the game is producing accurate statistics.

There are a lot of issues with how the game handles defense, player development, aging and player creation. They get tweaked and adjusted every year. Is there a problem with the minimum threshold for 2B? Maybe, it's something worth exploring further. Should the way players are rated be changed to better represent skills? Probably, there are several ratings which don't actually mean what one intuitively thinks they do.

If someone really wants to dig into this fangraphs has detailed double play data, but you would need to write a program to parse OOTP game logs to get the dp started/turned numbers. (Maybe Matt can add those as stats in the future )

Finally, the use of the 1-5 scale is detrimental to these discussions. With relative ratings on, the 20-80 gives an actual average and the standard deviations from the mean. Also for simming use a full season so numbers are easily comparable.

[Furious]

Someone brought up Bill James. In one of James’ books he discusses irrational skepticism.

There’s rational skepticism, where you doubt the veracity of something because it sounds false based on what you already know (or at least think you know), and then there’s irrational skepticism, where even in the face of overwhelming evidence you cling to excuses to disbelieve something because the truth is particularly unpalatable to you.

I’m not accusing anyone of anything; I don’t know what’s inside anyone’s head. I do think some of the objections that are being raised are starting to sound like irrational skepticism.

When Isryion told me about his Turn DP 1 second baseman who could turn DPs almost with the best of ’em, I was skeptical as hell. I thought of some of the same things some of you did: a) small sample size (he had, I think only a couple of seasons of data, and I believe they were only partial seasons); b) dissimilar DP opportunities when compared to the other 2Bs in the league (was it a pitching staff that allowed a lot of baserunners, or relatively few? Was it a ground ball staff or a fly ball staff? Was it a high-strikeout or a low-strikeout staff?)

One of the wonderful things about OOTP is you can use it to test things under conditions you could never replicate in real life. I can’t find six teams of exactly identical rosters in real life, but I can make them in OOTP. I can’t run those teams through ten seasons under identical conditions without all the players aging ten years in real life, but I can in OOTP. I can’t make sure none of the players get injured in real life, etc. OOTP allows us to conduct experiments that would be impossible to conduct in real life.

Such experiments may not be particularly useful for proving real-life hypotheses, but they’re great for testing how things work in OOTP.

Remember, I went into this experiment as a skeptic. I didnt think I was going to find strong evidence that the Turn DP rating has minimal effect on DPs, and I didn’t want to find evidence that the Turn DP rating has minimal effect on DPs. I found that evidence anyway, and I’m not interested in burying my head in the sand. I love OOTP, and I want it to be as realistic as they’re able to make it. When it looks like something’s wrong with it, I want them to fix it, not pretend the problem isn’t there.

Does anyone really think that a guy with a 30 Turn DP rating should be 90% as effective at turning DP’s as a guy with a 140 rating? Or do think that “just happened” because I didn’t set the study up exactly the way you think I should have? Are you sure you’re not just indulging in irrational skepticism?

[Isryion]

Originally Posted by Garlon

It matters if a player is a 2B or not because they also have an "experience" rating in the editor. Try putting a player there with 0 experience at 2B for a season with that minimum DP rating and see what happens.

This is exactly what I did (and mentioned above) and was cause for Furious's experiment. I took an elite 3B who had no 2B rating, with a DP rating of 1 and moved him to 2B. There was no noticeable difference in his ability to turn DP before or after he gained experience and he managed to turn DP at the rate of similar 2B throughout the league.

[Isryion]

Originally Posted by Ktulu

If you have the skills to play the position at all you have a baseline ability to turn double plays. The rating is there to help distribute DPs appropriately between the skill levels of major league players. It's not a representation of a player's physical ability to make the pivot at second base. You can easily see this by looking at the correlation of DPs by 3B and their turn DP rating.

See my above post, though I'm not sure what you're getting at exactly in this statement. This was a player we all assumed was a 3B and 3B only who I moved to 2B.

The reason 3B don't turn as many DPs in OOTP isn't due to their rating, it's that most of their innings are at 3B and they don't have as many opportunities.

[Isryion]

I'll just wrap up my thoughts with this, and a lot of it echoes what Furious is getting at in his post.

I personally think it's pretty clear to me the DP rating doesn't do much. My own experiment in the league and Furious's numbers confirm that. Is it worthless? Not quite, but a minimum rating isn't a significant difference between an elite rating. I think that works very different from every other rating in the game. I think a minimal double play rating should pretty much be bad enough that it prevents a player from being a reasonable choice to play a middle infield positions.

Now I think there are three camps based on the responses. The first is that you agree there is an issue with the rating and it should be changed in some way. The others are:

1) You still think the DP rating matters more than the tests show and these results don't satisfy you. I'd strongly urge you to run your own test. Eliminate the questions you see in the results. For what it's worth, I ran my own tests where I put 1B at 2B for a season and they were able to turn over 100 DP a season despite never having played 2B and having minimal DP ratings, some of them with OF or other 1B playing SS.

2)You think the DP rating is acting as intended and that turning DPs is easy and the range between major league baseball players regardless of other skills is small. These results show that OOTP simulates that and nothing needs to be changed.