Professor Luckymann’s Lab NeL Project

[luckymann]

[luckymann]

[luckymann]

Funny how often all it takes is a different perspective from the one you previously held to move the needle on long-standing roadblocks.

After much futility, starting this project has inspired in me what I think is a solution to my problem with giving the pitchers a similar career trajectory to position players.

Sadly unlike for the hitters, where one could change the stats in the DB, there's no way this methodology could ever be anything other than a manual application, but still for my purposes it's a major breakthrough - or, at least, I hope it is!

So I'll now be reapplying individual pitch, control and movement ratings to pitching candidates in this study each season.

Here, by way of example, is Harry Buckner's career trajectory in these metrics:



Anyway, just thought those of you following along would be interested in hearing about this. Let's see if it is actually a workaround or mere folly on my part.

On to 1902!

G

[luckymann]

[luckymann]

[No Pepper]

I'll be following along! I'm curious to know your roster sizes (active and reserve) and where these players (control sim) are on the depth charts vs how much playtime other reserve players are getting during the season. With injuries and trading low and finances off, there's only so many chances to play. It's clear that your treated players are getting significant starting time on active rosters. I suspect that it won't be until around 1920 that control players will start to crack regular lineups on the whole, since that is when the '24 database declares the real life statistical records as "major league". Keep up the good work, maybe Hill will break the mold first.

[luckymann]

Originally Posted by No Pepper

I'll be following along! I'm curious to know your roster sizes (active and reserve) and where these players (control sim) are on the depth charts vs how much playtime other reserve players are getting during the season. With injuries and trading low and finances off, there's only so many chances to play. It's clear that your treated players are getting significant starting time on active rosters. I suspect that it won't be until around 1920 that control players will start to crack regular lineups on the whole, since that is when the '24 database declares the real life statistical records as "major league". Keep up the good work, maybe Hill will break the mold first.

Hey bud, nice to have you aboard! Hope you've been keeping well.

Rosters are all at default and will change as per the game auto-evolution and -expansion. So currently they are at 25 for Active and unlimited Reserve. Obviously these guys will start getting more game-time when the minors kick in (1912?) but until then as you've seen very few of them are even seeing time on the AR.

Injuries and trading are tricky. I do plan to tick both up in a staggered fashion starting in 1905, but it's a double-edged sword for both.

Clubs usually trade for better players - especially with no financial motivation here with that element of the game not being applied - which can harm the lesser players' chances of increased game-time as much as help. Same with injuries, the candidates can get injured as well.

I will definitely be keeping long-term injuries low so guys don't miss chunks of time.

I'm trying to manage a subjective-objective balance here as well, at least in the Control sim. On the one hand I am trying to see for myself how these players are treated under default conditions, but then again I think I already know the answer to that (worse than they should be) and am trying to prove it, or at least let the game prove it for me. Which is why I am reluctant to alter settings.

The purpose of the Treatment sim is more personal. I mainly want to see how my curated edits are performing and will be using the results from this exercise to further fine-tune them. I know they aren't perfect (I reckon I've over-compensated a tad, especially with pitchers) and honestly don't believe "perfect" exists in this situation, so every time I put them through their paces I get more of an idea what needs to be done to at least keep improving them. Thus why you see them feature in one shape or form in pretty much every one of my (many) saves. The Everyman League, for example, has them edited on import but is a dev-only save, which lets me see how they behave under those conditions.

I'm not arrogant enough to think that my methodology is a viable candidate for official OOTP usage or anything but I do hope it at least encourages the devs and higher-ups to think outside the box for these guys and not be beholden to raw stats, which for the majority aren't representative of their abilities and it would seem aren't working in-game as they should.

Keen to have you, or anyone following along, keep firing thoughts, tips, feedback and requests and I'll happily try to accommodate them as much as I can, either during the sims themselves or in the post-analysis part.

Tracking usage is a tough one, however, as I am trying to get this done in a timely fashion and that means fairly zipping through the seasons themselves. I'll certainly now be keeping more of an eye on this and will include what I do manage to glean in my annual recaps.

Pete comes in with 20 CUR - 54 POT ratings and the issue I'm seeing with a lot of these early players may well be the case with him as well - that first recalc just kills them and their POT plummets while their CUR barely moves at all and they end up being like 25-25 guys, not MLB material. Let's hope he does indeed break that pattern. He's one of my personal faves from these formative years.

Cheers bud, appreciate the support for something that is really important to me.

G

[luckymann]

[luckymann]

My apologies for the constant tinkering with how the stats are presented - struggling to find the best way to do so informatively while also presenting well. Think I'm nearly there with the latest changes.

G

[Charlie Hough]

These seem like promising results thus far. Personally, I'd like to test the solution that I recommended in the bugs forum, where we apply a simple modifier to NeL player stats, to multiply them by the right number, to make them equivalent to the same stats compiled over 154 or 162 games, etc. But I don't think it's possible to mass-edit the OOTP database as easily as we could in past versions. EDIT: Also, as you've pointed out, the issues with early 1900s players have a different cause than what's impacting NeL players when OOTP starts importing them in the 1920s.

I think your experiment shows that, however we achieve the right modifications to the base stats for the right players and context, it's going to massively improve results. So, hopefully we're going to see some kind of fix in OOTP25.

[luckymann]

Originally Posted by Charlie Hough

These seem like promising results thus far. Personally, I'd like to test the solution that I recommended in the bugs forum, where we apply a simple modifier to NeL player stats, to multiply them by the right number, to make them equivalent to the same stats compiled over 154 or 162 games, etc. But I don't think it's possible to mass-edit the OOTP database as easily as we could in past versions.

Maybe I'm wrong about that and someone can steer me in the right direction. But I think your experiment shows that, however we achieve the right modifications to the base stats, it's going to massively improve results. My proposed solution is based on what would be relatively easy to code and still provide a very effective fix. Your method is a slightly more sophisticated approach that relies on third-party's MLEs, but it's still fairly simple. So, hopefully we're going to see some kind of fix in OOTP25.

I don't see what you're suggesting as a workable solution for a number of reasons. These guys aren't getting caned because of the "make bad" discounts (although I admit there would be some punitive downgrade to their stats if they have very few AB), they are getting caned because of the level assigned to the leagues they played in. Multiplying the number of games out won't do much to help them other than remove that fairly inconsequential discount. Plus, some of these guys did only have a handful of games some seasons so it is imperative that the discounts remain in place for them, so how would you separate the two categories in an equitable manner so that Ballplayer Williams who went 6-for-10 with 2 homers doesn't have that turned into a monster season?

It also comes down to the old adage: dirty in, dirty out; many of the issues are being caused because the DB is littered with erroneous stats and nothing will be fixed until that issue is addressed as well.

I might be wrong. It would be easy enough for you to test. Just pick a spread of players as candidates from 1900-1950 and start as I have, then make the changes you are suggesting in the editor as they come into the league and after the first recalc.

You'd only need to do a handful of candidates, so not much work involved.

Might be worth your while to do such a test so you can push your argument with more confidence but I think you'll find your logic is awry.

For mine, the only true solution that would possibly work across the NeL board is what I call "season strength modifiers", whereby each season of all the various Negro and Latin leagues are given a rating that adjusts the stats from that season in toto. Even this has holes I haven't been able to plug and I have given this matter more thought than just about anyone, I'd say.

The quest for the Holy Grail continues.

G

[Charlie Hough]

Okay, I see what you're talking about here. If the leagues are classified as BC, A or something like that, then player ratings will suffer. It's an issue of how all those early NeLs are classified or NeL players are evaluated vs. those from the 1920s onward.

Yes, my solution isn't going to resolve that. I'm only focused on the problems with player ratings for the NeL players that OOTP imports starting in the mid-1920s. Those are the imports that are affecting the most users, since people are using that option in the game and not starting in the early 1900s. Those issues are also drawing the main accusations of unfairness and disrespect, due to the high visibility. Obviously, there's also an issue with players from earlier decades, but, as you're pointing out, that has a different cause in the sim.

Fortunately, I think the problems with the 1920s onward are easy to fix. Yours are harder to address, and my understanding is that there isn't widespread agreement among historians on how to classify those early leagues or evaluate player performances within them. But it would be good to see something done to help improve the situation. In OOTP, if players have really outstanding stats in historical minors, they'll typically receive average to great ratings. I see that all the time, especially if they compiled them at the AAA or AA level. But, if they compiled them at lower levels, then they'll be adjusted downward, and that's probably the issue you're trying to resolve.

[luckymann]

Yeah, it's an incredibly difficult problem because some of those early leagues weren't great and the stats are sketchy. I see no real wholesale solution for all NeLers across the board. Hard as it is to admit, I think a lot of the lesser players have to be forgone and the changes made on an individual player-by-player basis. Kind of like segregation all over again, which I don't like one bit, but it seems the only way to have the top-flight guys perform as they should.

That's why, no matter the results of this study, I am not taking pot-shots at the devs. Yes, I'm extremely frustrated at the seemingly arctic pace of progress but I understand it should above all be done right. All I'm trying to show is how it looks under one such methodology, imperfect but hopefully an improvement. It's all about moving the needle in the right direction at every opportunity.

G

[Charlie Hough]

Yes, I wouldn't hold out much hope for OOTP Developments addressing any of the early players unless it can use a third-party source as a "database" of stats for importing and creating specific players. That might be what you're ultimately trying to build.

In the past, we've seen support for a couple of databases built by the community, so if you could create something like that for select players, with historical IDs to be matched up and stats to be used instead of the default stats in the historical and historical minors databases, then maybe the developers would entertain the idea.

Unfortunately, this issue is on the more obscure end of NeL representation, but if you can complete your tests and build a database, then maybe your fix could be paired with a fix for the adjust/weaken issue with the NeLs from the 1920s onward. My hope is that we might get an adjust/weaken fix for OOTP25. If that can be paired with a fix for select early players, then it becomes an even better improvement, and it's one that has good PR value for OOTP Developments.

[luckymann]

Originally Posted by Charlie Hough

Yes, I wouldn't hold out much hope for OOTP Developments addressing any of the early players unless it can use a third-party source as a "database" of stats for importing and creating specific players. That might be what you're ultimately trying to build.

In the past, we've seen support for a couple of databases built by the community, so if you could create something like that for select players, with historical IDs to be matched up and stats to be used instead of the default stats in the historical and historical minors databases, then maybe the developers would entertain the idea.

Unfortunately, this issue is on the more obscure end of NeL representation, but if you can complete your tests and build a database, then maybe your fix could be paired with a fix for the adjust/weaken issue with the NeLs from the 1920s onward. My hope is that we might get an adjust/weaken fix for OOTP25. If that can be paired with a fix for select early players, then it becomes an even better improvement, and it's one that has good PR value for OOTP Developments.

The one area, Charlie, where something like what you're suggesting would definitely improve is proneness. Because the game sees so few GP for these guys, it gives the majority of them a FRAGILE rating with 160 across the board.

Not so much of an issue with injuries set at the low end, but all contingencies need to be covered so this is something that needs addressing in one way or another.

G

[luckymann]

[luckymann]

[luckymann]

[No Pepper]

Taking a look at Pete Hill under the hood in the OOTP24 database, it is easy to see why he's "struggling" to get playing time in your Control Test. His real life stats are designated "Single A level" for the majority of his career from 1904-1908. You saw a bump in ratings in 1905 as expected (Triple AAA stats).

It is interesting that the 1907-1908 NACBP that Hill participated in is given Major League status in the database. He won't see that level of ratings bump until 1920 but his recalc during the interim will roller coaster from Single A to Triple A. It will be interesting to see how his ratings fluctuate during this time.

[luckymann]

Originally Posted by No Pepper

Taking a look at Pete Hill under the hood in the OOTP24 database, it is easy to see why he's "struggling" to get playing time in your Control Test. His real life stats are designated "Single A level" for the majority of his career from 1904-1908. You saw a bump in ratings in 1905 as expected (Triple AAA stats).

It is interesting that the 1907-1908 NACBP that Hill participated in is given Major League status in the database. He won't see that level of ratings bump until 1920 but his recalc during the interim will roller coaster from Single A to Triple A. It will be interesting to see how his ratings fluctuate during this time.

Yeah, same with most of the guys from the pre-1920 era. Some penalty is fair enough, but not single A.

I've been advocating a more granular approach for these guys, applying some sort of "League Strength" modifier for the various Negro and Latin leagues to players' stats from each season. That's where I'm heading next with my edits in what should be the final step for them.

Mine already take some of this into consideration, given I split the raw stats with the MLE and Eric applies this discount where necessary into his calculations. You see it in these earlier guys more than the later ones.

Here, by way of example, are the lines in my calc for Pete:




That's using EC's v2.0 MLEs, where he upgraded most players - the difference was far greater in the earlier version.




The issue I have with my own workings is that I'm worried I am overcorrecting, especially now with Eric having moved his workings upward.

I have Pete maxing out at a 360 AVG with 9 HR in his age-31 season. I don't think that's outrageous, especially given the game will always drag him down to that MiLB-level imprint over the course of a season.

All the same, I'm going to look at the results from this study and make a call on that. If so, I'll change my methodology somewhat to reflect it. It may simply need a bit of fine tuning so their peak seasons aren't so peak, because I think I've got things pretty close to where they should be career-wise in the OOTP context.

We shall see.

G