Historical Mode updates

[fredbeene]

you mean we still don't have real coaches, managers, gms?

[fredbeene]

are the additional picture slots available so you can have and use different pics of ball players (ie face, batting, fielding)

[mgom27]

Originally Posted by fredbeene

you mean we still don't have real coaches, managers, gms?

No Insiders,Announcers and so on.

[Hoiles]

Originally Posted by Syd Thrift

I understand why it was done but tbh the BABIP thing feels like a sop to people who don’t want to accept that for most pitchers it’s not a thing they can control and that guys like Allan Anderson and Johnny Podres just got hit lucky for a season or two (which, even though he was practically the poster child for the anti-DIPS crowd back in the day, his historical BABIP was below average for his career; he just had one really lucky year where opponents BABIPed .250). I know the mantra is always “just turn it off” but OOTP isn’t Strat and I think it doesn’t do super well when it tries to be Strat.

I think it could work well on a limited basis. Some historical pitchers like Warren Spahn and Jim Palmer underperform with the BABIP model. Whether or not it’s only because they had great fielders behind them or due to an additional “pitch to contact” skill is anyone’s guess. The engine would need to be sophisticated enough where Palmer isn’t doubly benefiting from this new rating along with the great D.

[Syd Thrift]

Originally Posted by Hoiles

I think it could work well on a limited basis. Some historical pitchers like Warren Spahn and Jim Palmer underperform with the BABIP model. Whether or not it’s only because they had great fielders behind them or due to an additional “pitch to contact” skill is anyone’s guess. The engine would need to be sophisticated enough where Palmer isn’t doubly benefiting from this new rating along with the great D.

So… Palmer I see, although I do think that IRL his boost was probably due to playing in front of a defense Earl Weaver had carefully constructed to be one of the greatest of all time. It helped that Brooks Robinson, a Hall of Famer, was out there, but he stuck with the weak hitting Mark Belanger for years because he was an 80 fielder, and he also ran some very high quality second basemen and center fielders as well. Spahn to be honest is about right in line with the thing where the best power pitchers get a 10-15 point boost in BABIP.

[Hoiles]

Originally Posted by Syd Thrift

So… Palmer I see, although I do think that IRL his boost was probably due to playing in front of a defense Earl Weaver had carefully constructed to be one of the greatest of all time. It helped that Brooks Robinson, a Hall of Famer, was out there, but he stuck with the weak hitting Mark Belanger for years because he was an 80 fielder, and he also ran some very high quality second basemen and center fielders as well. Spahn to be honest is about right in line with the thing where the best power pitchers get a 10-15 point boost in BABIP.

Other Orioles, like Cuellar and McNally, also had ERA/FIP differences, though not to the same extent as Palmer. One might also say this new rating represents an additional "Weaver effect" that isn't present in the current engine. It will be interesting to see how this stuff is implemented in the game.

[Garlon]

Matt, can I respond to some of the things in this thread or should we wait until release?

[Matt Arnold]

Originally Posted by Garlon

Matt, can I respond to some of the things in this thread or should we wait until release?

Some of the stuff is still in fluxx depending on how much of your hard work we can incorporate in

But overall, I wish we had a better way to adjust to some of the historicals of the past, but in the end, this really felt like the best way to give people credit for what they did. If you are playing a replay, then this will significantly help the best seasons be the best seasons. And if you're less interested in a true replay, then you probably have either 3 year or career recalc, which won't have those outlier seasons at play.

And it's certainly the sort of thing we will keep an eye on to make sure it's doing what is intended, and that the majority of the play modes are handling it in the correct way.

[Garlon]

Mat, they asked some questions about Jim Palmer and Warren Spahn that I can provide more information on for those that are interested.

[Matt Arnold]

Originally Posted by Garlon

Mat, they asked some questions about Jim Palmer and Warren Spahn that I can provide more information on for those that are interested.

Yeah, I mean obviously any general information you learned in your investigations, you're always welcome to share

[Garlon]

Jim Palmer had a career .251 BABIP, while the league had a .280 BABIP during those seasons. How much was Palmer and how much was the great defense of the Orioles in those seasons? When we factor out team defense and ballpark factors, Palmer adjusts to .268 BABIP for his career. There is a difference of 29 points between the league average of .280 and the .251 that Palmer had and 12 points of this was something Palmer was doing, and the other 17 points was the defense from his team. Palmer was about 350 hits saved combined with his ability and the defense from his team. If you put him behind an average defense he should save about 135 hits with his BABIP ability.

Warren Spahn had a career.261 BABIP against, while the league had a .294 BABIP during those seasons. When we adjust for team defense and ballpark factors, Spahn ends up with a .268 BABIP. There is a difference of 33 points between what Spahn did and the league average, and in this instance 26 points are something Spahn was doing, and 7 points were from the defense of his team.

Hopefully, we will get much better results in OOTP for some of these pitchers. Spahn does always underperform in OOTP. That difference of 33 points of BABIP results in saving over 500 hits allowed across his career, so that should have a very positive effect on his performance. Even playing behind an average defense he should still save about 375 hits over those seasons.

[Syd Thrift]

According to BBRef the difference with Spahn is .261 vs .276 (and .294 would be insanely high for the 50s and 60s). Like I said, that's right in line with the bump that real-life power pitchers get. I'm not sure that the game gives power pitchers the same 10-15 point boost (probably 15 points would be the maximum) but that kind of does explain the difference.

My dog in this fight by the way as a fictional player is that if top power pitchers don't get that boost, they will underperform in fictional leagues as much as real-life ones. Otherwise, if there's a demonstrable effect, it really should be tied into something other than "BABIP effect", which TBH reeks of overfitting - and like overfitting when you do regression analysis, it won't do anything for future returns, so to speak.

Palmer to me demonstrates one of two things:

1. The effect of fielding may not be high enough (although TBH my general aesthetic/intuitive sense it is pretty pronounced and pretty high already).

2. Historicals don't do a super great job of assigning super high ratings to individual fielders. This, at least, is something I see borne out by other complaints - people playing the 80s complain that Ozzie Smith never gets 80s, I'm sure that people playing the 70s say the same thing about Brooks Robinson and Mark Belanger, and so on. I appreciate that it is really hard to suss that out from basic fielding stats but maybe whatever happens right now, maybe it needs to grade players on a curve WRT fielding ratings, so to speak, so that the top, say, 2-3% of players at a position get 80 range, etc.

I'm sure that the latter makes some people who do the analytics feel icky but the only other alternative I can really think of is to go in and manually create/alter those ratings for players and that isn't something OOTP has ever really wanted to do in the past.

[Garlon]

You are right about the league average BABIP in the NL for Spahn. For some reason it is not accurately reported on the player page if you go to Pitching Advanced Stats>Pitching Against. For Spahn it shows that he was .261 while the MLB average was .294

This makes no difference though because we do have the correct BABIP from the era_stats file and Spahn does have the correct BABIP against in the pitching file.

[Syd Thrift]

OK but which one of us was right about that isn’t super important to me TBH. What I’m wondering is:

- Does OOTP have a Stuff-based modifier to BABIP, based on real life effects? (again it’s not necessarily high but you’ll see a similar effect on other big time power pitchers in MLB history, for example Nolan Ryan)

- If not, can this, way more than the “BABIP effect”, something that can be added to the engine for all leagues, not just historical ones?

- What about fielding? Has work been done to ensure than the Ozzie Smiths and Mark Belangers of the world get 80s, etc? This to me isn’t as huge of an issue because, selfishly, I play fictional and IME the fictional engine does fine at giving out 80 ranges and generating players who produce 2 or 3 wins worth of value with their glove, but IMO the best fix to a guy like Palmer is still going to be to start with fixing the fielding and then seeing how far that gets you in his performance.

The fielding metrics I’ve seen rank Brooks Robinson as far and away the best fielding 3B of all time and Mark Belanger a touch ahead of Ozzie Smith as the best of all time at his position. And on top of that, Bobby Grich is rated as a top 20 guy. Al Bumbry was also an extremely good defensive player if not on Jim Edmonds or Andruw Jones levels of dominance. Palmer is a big outlier but so was his defense.

[Hoiles]

Thanks for the responses, good to see OOTP isn’t resting on its laurels. The pitching model is fine as is but a tweak here or there can’t hurt.

As for the fielding imports, i never had any complaints with the year-by-year or 3 years ratings. When using career totals, some of the players who had extra long careers tend to have their fielding ratings neutered from including 35+ seasons (eg Griffey has bad ratings). I don’t think it’s a big deal as we can just use the other options.

[Garlon]

Any effect that a strikeout pitcher had in their BABIP will already be in their historical BABIP value. There is no need to add an additional boost on top of that. Whether or not the game does this for fictional leagues, I do not know.

For fielding ratings, players need to be rated relative to the league. You cannot just take the best defensive SS in any given season and automatically make them an 80/80 because the top defensive SS one season might make 5% more plays than average and, in another season, the best SS might make 9% more plays than average. These players cannot both be 80 simply on the basis that they were both the best SS in their league in a given season or the fielding results will not turn out properly.

Sometimes the best defensive player at a position was not that much better than average. Sometimes they had historic defensive seasons.

I have Al Bumbry as the best defensive CF in the AL in 1980.

For Jim Edmonds, I have him catching 4.3% more flyballs than the average.

[Garlon]

Regarding Griffey, you can take a look at his Sabermetric Fielding on BBR and see that as a CF he was -55 defensive runs from fielding plays and +60 defensive runs from Arm plays.

Everyone keeps saying that Griffey was a terrific defensive CF, but I do not see it in any of his statistics when it comes to tracking down flyballs. His Sabermetric Fielding only has him above 10 runs saved from fielding flyballs in 1995 and 1996. From 1989 to 1992 he is a combined - 11 runs saved from fielding plays. In 1998 and 1999 he is -33 runs saved from fielding plays.

I do not see any great defensive performance relative to the league with him in my adjustments nor does he turn out to be great in sabermetric evaluations.


He won 10 gold gloves from 1990 through 1999.

These are the best CF in the AL each season during that streak of gold gloves according to BBR:

1990: Roberto Kelly

1991: Lance Johnson and Devon White tie

1992: Devon White

1993: Lance Johnson, Devon White, and Kenny Lofton tie

1994: Kenny Lofton

1995: Griffey Jr. and Bernie Williams tie

1996: Griffey Jr.

1997: Darren Bragg

1998: Kenny Lofton

1999: Chris Singleton

There were other great defensive CF in the AL during those seasons and when he moved to the NL Andruw Jones was playing CF at a level that is greater than any other CF in history.

[Syd Thrift]

Originally Posted by Garlon

Any effect that a strikeout pitcher had in their BABIP will already be in their historical BABIP value. There is no need to add an additional boost on top of that. Whether or not the game does this for fictional leagues, I do not know.

For fielding ratings, players need to be rated relative to the league. You cannot just take the best defensive SS in any given season and automatically make them an 80/80 because the top defensive SS one season might make 5% more plays than average and, in another season, the best SS might make 9% more plays than average. These players cannot both be 80 simply on the basis that they were both the best SS in their league in a given season or the fielding results will not turn out properly.

Sometimes the best defensive player at a position was not that much better than average. Sometimes they had historic defensive seasons.

I have Al Bumbry as the best defensive CF in the AL in 1980.

For Jim Edmonds, I have him catching 4.3% more flyballs than the average.

OK so no offense but I’m beginning to feel like to some extent this is papering over a flaw in the game (there not being a stuff based modifier to BABIP - again, if I’m wrong in saying that, I’d love to be corrected). I appreciate that some will like that Warren Spahn will perform to his numbers but it more than bothers me that fictional Bloren Blahn with the same exact generated ratings will not, at least given that we know about pitchers with high K rates and BABIP. Also Tom Seaver and Roger Clemens (weirdly not Randy Johnson, although I have to say that none of the high K guys I happened to jump around and look at had *worse* BABIPs than average) (and I’d love to find a/ the study where I read this - it might be part of Tom Tango’s dive into it - but I can’t).

I’ll concede Palmer for now because again it doesn’t matter a lot to me, but I’m not sure that I buy that the fielding numbers “would be off” unless/until that’s actually tested. Also if this really is just a thing that we’ve decided that some pitchers have, why wouldn’t fictional pitchers have it then as well?

[Garlon]

I did say that for fictional leagues I do not know how the games handles any possible correlation between Stuff and BABIP. Matt may have some insight on that though.

Fielding ratings are very small margins and the game is calibrated so that a certain combination of Range and Arm values at SS will be a 1.000 relative to league in terms of making ranged fielding plays, and another combination of Range and Arm gets you to 1.002, and another for 1.004, and another for 1.075. In this way, you cannot just give a SS some extreme rating because it needs to correspond to the correct relative range rating of the player.

[Matt Arnold]

Originally Posted by Syd Thrift

OK so no offense but I’m beginning to feel like to some extent this is papering over a flaw in the game (there not being a stuff based modifier to BABIP - again, if I’m wrong in saying that, I’d love to be corrected). I appreciate that some will like that Warren Spahn will perform to his numbers but it more than bothers me that fictional Bloren Blahn with the same exact generated ratings will not, at least given that we know about pitchers with high K rates and BABIP. Also Tom Seaver and Roger Clemens (weirdly not Randy Johnson, although I have to say that none of the high K guys I happened to jump around and look at had *worse* BABIPs than average) (and I’d love to find a/ the study where I read this - it might be part of Tom Tango’s dive into it - but I can’t).

I’ll concede Palmer for now because again it doesn’t matter a lot to me, but I’m not sure that I buy that the fielding numbers “would be off” unless/until that’s actually tested. Also if this really is just a thing that we’ve decided that some pitchers have, why wouldn’t fictional pitchers have it then as well?

Fictional/modern players do have some slight tweaks to their babip yes based on factors like that. Those players it's based more on their repertoire, and it will vary, although it does tend to stay in a narrower range than how we are assigning the historical values.

We do have the capacity to assign this for anyone, and certainly will look to see how we can adapt this moving forward. I do think generally speaking in fictional leagues we will only want to set this for certain players if we want to create the next Spahn or Ryan or Rivera, I would imagine the majority of players we should be fine with keeping them in the defaults. But certainly it's something we will look at and see what changes might be necessary over the next months and years.