The Turn DP rating—does it do anything?

[Garlon]

You need to account for opportunities. The opportunities are affected by runners reaching 1B, by groundball rate, and strikeout rate. If you have player on 1B with 1 out and there is a strikeout, then you really did not fail at turning a double play because there was no opportunity.

For Jose Altuve in 2019 when you account for these things he was expected to turn 52 double plays and he actually turned 49 double plays, so he was -3 for double plays. You could also say he was 49/52 = 0.942, which is about 5.8% fewer double plays than average.

For Yolmer Sanchez, he had 108 double plays, but he was only expected to have 93, so he was 15 double plays above average. This is 108/93 = 1.16, so he was 16% better than average.

For Eric Sogard he had 28 double plays and was expected to have 28 double plays.

The best in 2019 was Kolten Wong with +18 double plays and the worst was Adam Frazier with -11 double plays for the season. That difference of 29 double plays between the best and worst is about 7 defensive runs for the season.

[Furious]

I just wanted to clarify something I wrote earlier. I spoke of "routine" double plays, ones that just about anybody could turn.

I should have been more specific. I think there are a lot of plays that the player the ball is hit to could make even if he is not particularly skilled. The second baseman on a 4-6-3 or the shortstop on a 6-4-3 doesn't have to be Superman if the ball is hit right to him.

The player who has to take the throw and fire to first base has a much tougher play, especially if he's playing second. I don't think that's a play that just any player could routinely make, and I think if you put Yadier Molina or someone else with no middle infield experience in that position, he would screw it up a lot.

But, as we've discussed, we have no data to support that, because real teams don't put players with poor double play skills at second base.

[quillenl]

Originally Posted by Garlon

You need to account for opportunities.

I agree but didn't see that data readily. Did you find the double play opportunities on Baseball-Reference or some other site? Kind of amazing to see Sogard have half the rate of double play opportunities as the league leaders.

[Furious]

Garlon—please consider starting your own thread. This one is supposed to be dedicated to a very specific issue—the effects of the Turn DP rating on fictional players with very low Turn DP ratings. You have posted very little that addresses that topic and you’ve categorically ignored attempts to guide you back to it.

You have a lot to say about double plays and fielding issues in OOTP and I’m sure a lot of it is interesting, but almost none of it has anything to do with my original post. Please take it elsewhere.

[mytreds]

Originally Posted by Furious

I don't think it is that easy. I think there are players who are really good at it and players who aren't very good at it. I would not assume all players have an adequate level of skill at something just because there are no statistics to prove they don't.

This is why as a scorekeeper you’re not allowed to assume double plays. It’s why any normally given error on a throw to first is not given (unless the errant throw allows the batter runner to advance) and the batter gets a FC.

[Rain King]

Originally Posted by Furious

Garlon—please consider starting your own thread. This one is supposed to be dedicated to a very specific issue—the effects of the Turn DP rating on fictional players with very low Turn DP ratings. You have posted very little that addresses that topic and you’ve categorically ignored attempts to guide you back to it.

You have a lot to say about double plays and fielding issues in OOTP and I’m sure a lot of it is interesting, but almost none of it has anything to do with my original post. Please take it elsewhere.

He's providing actual context, which you (by your own admission) haven't been able to do.

You have provided ample evidence that something might be off (and almost certainly is on the extreme), but you haven't been able to provide any kind of number OOTP should be shooting for (which is what Garlon is actually doing).

I understand you would prefer the topic stay on your very specific case, but any adjustments to the engine shouldn't just consider your case.

I, for one, greatly appreciate your bringing up this discussion but I do wish you would be a little more open to some of the skepticism to the conclusion you came in having already decided.

[Furious]

Originally Posted by Rain King

He's providing actual context, which you (by your own admission) haven't been able to do.

Look, unless someone can find me statistics from an instance when a real-life manager took the guy with the worst-double-play-turning skills on the team and made him his starting second baseman for several seasons, that kind of context just doesn’t exist. We’re trying to measure something that can happen in OOTP but which has no real-life parallel. Paragraph after paragraph analyzing Bill Mazeroski and Nap Lajoie isn’t helpful. It’s off-topic.

Allow me to contrast one of the people I have been disagreeing with, kriskolic, with Garlon. kriskolic has remained on topic (with one exception—when he quoted Garlon!). His assertion that double plays are relatively easy for any player to turn is something I disagree with, but it is central to the topic. We’ve gone back and forth on that a bit, and it seems (to me) like we’ve reached the point where there’s nowhere else to go on it without repeating ourselves. But I don’t object to kriskolic bringing it up and making his argument.

Anyway, my intent is not to bash Garlon; I’m just trying to keep the focus on the topic.

I understand you would prefer the topic stay on your very specific case, but any adjustments to the engine shouldn't just consider your case.

I am not unmindful of that. If someone is saying “fixing this problem for fictional leagues will screw up historical leagues”, then I’ll admit that’s an important consideration. I’m uncertain if that’s what anybody is saying.

I, for one, greatly appreciate your bringing up this discussion but I do wish you would be a little more open to some of the skepticism to the conclusion you came in having already decided.

Well, I did reach a conclusion prior to starting the thread. I didn’t start the thread in complete bewilderment, asking for opinions on how to interpret the results of the study I ran. I set up a test, was very careful to consider all the parameters I thought were important to consider, and said “look, I ran this test, and this is what I think it proves.” I’m open to hearing dissenting opinions. I hope you're not equating my disagreeing with those opinions—and trying to sort through relevant and non-relevant arguments—with refusing to consider them.

[OutS|der]

Has anyone run these tests while changing the league total modifiers?

Double up the league totals for double plays and then run your tests, see how much of a difference there is then.

[Furious]

Originally Posted by OutS|der

Has anyone run these tests while changing the league total modifiers?

Double up the league totals for double plays and then run your tests, see how much of a difference there is then.

I could do that, but what would be the point of it? I used the default LTM settings; most leagues use the default settings, or something close to them, if they're not using historical settings. I don't understand what the purpose of that test would be.

[Isryion]

Originally Posted by Rain King

I, for one, greatly appreciate your bringing up this discussion but I do wish you would be a little more open to some of the skepticism to the conclusion you came in having already decided.

To be fair, I'm the one who came to this conclusion long before the thread was up. Furious has actually been skeptical of it for some time, until he ran the tests he describes in the first post.

Almost all the questions/criticisms are ones that he and I had already discussed when I first told him what I thought.

I ran a test that responded to one question that I felt so far could be easily answered (non-2B turning DPs). I have zero problems with the questions but I don't understand why people aren't running tests that provide data or clear up issues that they think are still out there. Is it because they're not testable? I'm really not sure what kind of test you could run at this point, besides repeating Furious's test and hoping to get different numbers, that would show the double play rating is more significant than we believe.

What questions, besides deciding base lines for absolute minimal double play ratings are left to answer?

[Isryion]

Originally Posted by quillenl

I agree but didn't see that data readily. Did you find the double play opportunities on Baseball-Reference or some other site? Kind of amazing to see Sogard have half the rate of double play opportunities as the league leaders.

He hasn't explained it. I asked awhile ago so I could try to apply it, but maybe he missed it. There's a Bill James article that shows how to calculate it. As I understand it -- the formula is based expected results are based on an average for the era. Then calculating how many opportunities they would have had relative to that number based on runners on first and then adjusting for ground balls.

We could be skeptical of how accurate that is for a lot of reasons but none of them really matter because ultimately they are measuring a league average for teams that put players that are, at the least, competent at turning the double play at 2B and SS.

In OOTP, you can put an incompetent player at a position where they need to turn a double play (part of the fun of the game, but not something a real major league team would do except in the rarest of circumstances) and they will do it at a competent rate. This is my central problem with this since we can't actually see these players play. Is there another rating in OOTP where you would see the minimum and expect a player to be competent (not necessarily average, but not a disaster) when performing a skill related to that task? I'm asking seriously because I can't think of one.

[OutS|der]

Originally Posted by Furious

I could do that, but what would be the point of it? I used the default LTM settings; most leagues use the default settings, or something close to them, if they're not using historical settings. I don't understand what the purpose of that test would be.

It could show the DP numbers are being held down by the league totals. The gap between best and worst would be much greater compared to lower totals where it wouldn’t be.

The purpose is to show you that the league totals play a large part in the gap between best and worst.

Just like with HR, make a player with 200+ power with 2020 totals and he’ll hit a lot more HR’s then someone with power of 1, yet change the HR totals to say 1994 and he’ll only hit half as many compared to the worst since they can’t go any lower.

[Furious]

Originally Posted by OutS|der

It could show the DP numbers are being held down by the league totals. The gap between best and worst would be much greater compared to lower totals where it wouldn’t be.

The purpose is to show you that the league totals play a large part in the gap between best and worst.

I don’t know why you think I need to be shown that. I’ve been messing around with LTMs for years. I know what they do.

Doubling the number of double plays just to see a larger gap between best and worst strikes me as a completely pointless exercise, but If it’s something you’d like to see, I’ll be happy to send you the file so you can run the test yourself. PM me your e-mail address and I’ll send it along.

[OutS|der]

Originally Posted by Furious

I don’t know why you think I need to be shown that. I’ve been messing around with LTMs for years. I know what they do.

Doubling the number of double plays just to see a larger gap between best and worst strikes me as a completely pointless exercise, but If it’s something you’d like to see, I’ll be happy to send you the file so you can run the test yourself. PM me your e-mail address and I’ll send it along.

Whatever then. Don’t accept other views. Stick to your narrow view from limited testing.
I’m done trying to offer help or suggestions here.

You want to ignore the affect of league totals then go ahead. Just shows how meaningless your testing is so enjoy arguing with yourself

[Isryion]

Originally Posted by OutS|der

Whatever then. DonÂ’t accept other views. Stick to your narrow view from limited testing.
IÂ’m done trying to offer help or suggestions here.

You want to ignore the affect of league totals then go ahead. Just shows how meaningless your testing is so enjoy arguing with yourself

I don't see the need for this kind of response. Furious said he'd run the numbers.

I think the reason the league totals aren't super useful here is that we're seeing these totals in a modern league already and the concern is at the low end of the scale, so lowering the league totals just shrinks/increases the scale we're talking about when the concern is the low end in modern leagues.

Doubling the scale would increase the number for both types of players and obviously increase the gap, but it would also create an unrealistic situation and I'm not clear how it would show these players with minimal rating aren't turning too many double plays. I think it's clear that changing the league totals would change the spread.

If you think a player with a minimal double play rating should be able to turn 100 DPs a season in modern baseball, and it's working as intended, that's totally fine and it's a valid perspective.

Edit: In your example you use POW between today and 1984. You mention the spread would be smaller (and I agree), but do you foresee the players with minimal HR hitting more than 1 or 2? Or, do you foresee a setting where players with minimal ratings for POW hit at least close to an "average" number of HR?

I'm asking seriously because I feel like that's how the DP rating is different here. Players performing competently despite not really having the skill. Even if we change the rating to 10x the modern values for HR, I'd expect to see players with minimal ratings hit very, very few home runs compared to other players and, even if you didn't know their ratings, it would be clear they are low.

[Furious]

Originally Posted by Craig Scarborough

To those who've participated in this discussion, a big thank you. I've enjoyed reading the back and forth on this topic.

It's pathetic that I even have to call this out, but to see an internet "discussion" not degrade into a barrage of bickering and name-calling has been refreshing.

Originally Posted by OutS|der

Whatever then. Don’t accept other views. Stick to your narrow view from limited testing.
I’m done trying to offer help or suggestions here.

You want to ignore the affect of league totals then go ahead. Just shows how meaningless your testing is so enjoy arguing with yourself

Took less than 24 hours!

[Craig Scarborough]

Originally Posted by Furious

Took less than 24 hours!

But we had some good times, didn't we?!

[Bobfather]

I apologize if this has already been answered. To me, the most important stat(s) of all is Wins and Losses. Was there any significant difference in wins, between the teams?
Also, where all the 2B rated the exact same offensively?
Thank you for all the hard work you have put into this.

[Furious]

Originally Posted by Bobfather

Also, where all the 2B rated the exact same offensively?

Yes. Aside from the two second basemen whose Turn DP ratings were changed, every team is exactly the same in every detail except the name of the team, and the middle initial of each player (which is also an identifier for which team they’re on).

To me, the most important stat(s) of all is Wins and Losses. Was there any significant difference in wins, between the teams?

Almost none. I’ve attached an Excel file showing the standings for each season, and the ten-year totals and averages.

Every team averaged between 68 and 72 wins over the ten seasons. Remember, this is in a 140-game season, so that’s two games under .500 to two games over .500.

Since changing the ratings didn’t even result in a big difference in double plays turned, it would be strange indeed if it did result in a big difference in wins and losses.

Even if there had been a big difference in double plays, I wouldn’t have expected it to effect the teams’ records much. As one or two people have pointed out here, the effect of double plays on expected runs is small. You need a lot of them to make a difference.

Also, while wins and losses are the name of the game, it’s dicey to try to use team records as proof of anything. Two teams can finish, say, ten games over .500, but one team might have a +100 or better run differential while the other might have a negative run differential. If one team—either one—turned 10 double plays more than the other, would you say that explains their records?

Also, consider these two oddities…

1. In the tests I ran, there was one starting 2B (who played for the “A’s”) who had a better Turn DP rating than the five other starting 2Bs in the league, and he did average a few more DPs per season than the other 2Bs. But his team actually had the second-worst aggregate record over the ten seasons. So would you say that those extra DPs actually hurt his team, or would you say that they probably just had so little effect that they weren’t really a factor either way?
2. There was one starting 2B (who played for the “F’s”) who had a much worse Turn DP rating than the five other starting 2Bs in the league, and he did average a few less DPs per season than the other 2Bs. His team did have the worst aggregate record over the ten seasons. But… that only happened because one season was a total outlier; in the “3c” season, the F’s finished 51-89. It was the only season in the ten years that any team won fewer than 61 games; the F’s actually had a winning aggregate record the other nine seasons. That being the case, would you say that the team was hurt by turning a few less DP’s per season, or that it likely just didn’t matter?

Thank you for all the hard work you have put into this.

You’re most welcome.

[brotherblues]

Have you guys run the same test in OOTP 22?