WBL Rule Change Thread

[Lloyd Carr]

The only issue with changing the way stars are handled, is it may make evaluation of players more straight forward. Which could be a disadvantage for people like Gooksta who have spreadsheets.

 

[OU11]

The only issue with changing the way stars are handled, is it may make evaluation of players more straight forward. Which could be a disadvantage for people like Gooksta who have spreadsheets.

It's still the AI though, which isnt absolute

 

[Gooksta]

See, I think most people can agree with this. Which makes me question why we should maintain the status quo for what the stars represent.

There's a reason we say STARRRRZZZ. I think we should do something to make the stars more relevant.

I think people should stop looking at the starz.. but I like the stars over the numbers visually. Lol

 

[Lloyd Carr]

It's still the AI though, which isnt absolute

Very true.

So gooksta is still okay.

 

[Yankee151]

To me it's a skill ceiling thing to me, I don't want a league where I can just look at a player and see he's 70 potential or 4.5 potential (like actually 4.5 potential now how it is now) without thinking, If a players overall rating (stars/numbers) is tied to his WAR then his other ratings don't really matter. The way it is now forces you to do some work to determine which of those 1.5/2.5 star guys are actually good. I don't really care for the fact that shortstops get like 4 stars automatically due to their position but it's not like we don't see it coming or anything.

I just see a 100% accurate stars/numbers system(in terms of no positional bias yadda yadda) as disrupting the trade market (You want to trade me a 40 overall for a 60 overall? Get lost) whereas as of right now different things are valued by different managers and it creates more unique ways of building teams

That being said all those settings do is change presentation, the players themselves are still the same at their cores, so perhaps differentation would remain.

 

[Lloyd Carr]

I think people should stop looking at the starz.. but I like the stars over the numbers visually. Lol

I use stars more as a "comfort" thing. I evaluate guys based off a few things, and if the stars align with my evaluation of those things, I feel more comfortable with my evaluation.

 

[Gooksta]

I'll just say this.. if a player has high defensive and base running ratings.. his hit ratings don't have to be as high to be of equal or greater value of a decent offensive player with bad defensive or base running ratings.. it's a give and take

It won't that much.. most people are pretty good at judging talent.. my spreadsheet is really just to try and find productive well rounded players in later rounds of the draft..and make sure I don't miss any decent prospects that I would typically overlook.

 

[OU11]

To me it's a skill ceiling thing to me, I don't want a league where I can just look at a player and see he's 70 potential or 4.5 potential (like actually 4.5 potential now how it is now) without thinking, If a players overall rating (stars/numbers) is tied to his WAR then his other ratings don't really matter. The way it is now forces you to do some work to determine which of those 1.5/2.5 star guys are actually good. I don't really care for the fact that shortstops get like 4 stars automatically due to their position but it's not like we don't see it coming or anything.

I just see a 100% accurate stars/numbers system(in terms of no positional bias yadda yadda) as disrupting the trade market (You want to trade me a 40 overall for a 60 overall? Get lost) whereas as of right now different things are valued by different managers and it creates more unique ways of building teams

That being said all those settings do is change presentation, the players themselves are still the same at their cores, so perhaps differentation would remain.

Oh it would. People would still value what they value and prospects would be graded more harshly in those early years that they dont produce. I'll check it out and see what happens with someone like my rookie OF who is a beast prospect but is playing shitty

 

[OU11]

Dani Gut is a 4/5 right now and overall.

With those AI settings he is a 1/5 based on position and 2/5 overall

Each are more indicative of his production so far this year.

 

[Yankee151]

This reminds me of the OOTP sim someone did with teams full of 1 MLB player, I think the team with the power hitter (Adam Dunn) wrecked the team with the twink defensive shortstop

http://www.fangraphs.com/not/dangerous-experiment-a-roster-of-25-adam-dunns/

 

[Orlando]

Starz are based off physical ability.. I think baserunning ability is a small part of the equation also

I think that depends on the position as well

 

[Gooksta]

I think that depends on the position as well

It does

 

[OU11]

See, I think those are the metrics that make the most sense for evaluating how to rate a guy. Stars make the most sense (to me at least) as an evaluation of a players overall abilities as a baseball player. You've got their ratings and stats, and the stars indicate how good of a player they are when all of that is meshed into a general observable evaluation of a player's merits.

Example:

Darrach and Dani have the same ratings (well, close enough), but their star situation is a lot different. And that bugs me, because Dani is rated higher (star wise), even though he's not playng (nor has be ever played) better than Darrach.

I think the metrics in your post are a more accurate style for what to use when evaluating the overall skill of a player. And when I think of stars, the overall skill of a player is what I envision those stars represent.

I bet if you used those metrics in that post, Darrach would have more stars than Dani. Or at least it would make more sense to me for Darrach to have more stars. Maybe Dani has more potential stars because he's younger and the scouts think he has more long term potential? But overall, it doesn't make sense for Dani to be higher (in stars) than Darrach.

I'm curious if changing those metrics would support that or not.

Using the AI settings in that pic, this is the situation for Luong and Dani

Based on position

Dani 1/3.5
Luong 4/3 (yes his current is over his potential because he's playing well above what his ratings say he should)

Overall
Dani 1.5/2.5
Luong 3.5/2.5

 

[Lloyd Carr]

I think those evaluations are little more accurate than what we're currently using.

As far as the weights go, I'd probably change the weights a little bit. Probably lower current year and raise previous year.

Maybe something closer to: 35, 30, 25, 10 (obviously I'm just spitballing)

 

[Orlando]

I hate change. Go back to how they were!

Seriously though, I'd prefer the overall to be compared against others in the same position not everyone.

 

[OU11]

I hate change. Go back to how they were!

Seriously though, I'd prefer the overall to be compared against others in the same position not everyone.

I do too, and using some stats mixed with ratings would make them more usable than they currently are

 

[Orlando]

Also I would probably put the least weight on the current year and higher on the last two years and probably equal weights on those.

 

[Lloyd Carr]

Also I would probably put the least weight on the current year and higher on the last two years and probably equal weights on those.

Actually, I think I agree with that too.

I think that makes a little more sense as to how to evaluate talent.

 

[OU11]

Would 40/20/25/15 be good? Two years ago is just a long time ago to put more weight into than the current year. Guys can fall off a cliff in a year which would make their overall skewed

 

[Gooksta]

This will slow-change how prospects develop too btw..

 

[OU11]

This will slow-change how prospects develop too btw..

It will change nothing about how prospects will develop btw.

 

[Lloyd Carr]

Would 40/20/25/15 be good? Two years ago is just a long time ago to put more weight into than the current year. Guys can fall off a cliff in a year which would make their overall skewed

Yeah, I'd be okay with that ratio.

 

[Gooksta]

High schoolers could be even riskier

 

[Gooksta]

It will change nothing about how prospects will develop btw.

Ok. Never mind than

 

[osick87]

I think those evaluations are little more accurate than what we're currently using.

As far as the weights go, I'd probably change the weights a little bit. Probably lower current year and raise previous year.

Maybe something closer to: 35, 30, 25, 10 (obviously I'm just spitballing)

I agree with raising the previous years since those are the established seasons. I wouldn't change or raise rating because then you're likely going to have some inflated potential ratings and they're not 100% accurate.

Things that would be interesting

25, 25, 25, 25- Treats everything equal. Probably would have older players high even though they are dying this season.
25, 30, 30, 15- Puts emphasis on the current. The Ratings being worth a quarter would hopefully make prospects not always look piss poor.
0,30,40,30- Prospects would get fucked by this I'm assuming but this would be the way to see how good people are at producing.

 

[Travis7401]

See, I think those are the metrics that make the most sense for evaluating how to rate a guy. Stars make the most sense (to me at least) as an evaluation of a players overall abilities as a baseball player. You've got their ratings and stats, and the stars indicate how good of a player they are when all of that is meshed into a general observable evaluation of a player's merits.

Example:

Darrach and Dani have the same ratings (well, close enough), but their star situation is a lot different. And that bugs me, because Dani is rated higher (star wise), even though he's not playng (nor has be ever played) better than Darrach.

I think the metrics in your post are a more accurate style for what to use when evaluating the overall skill of a player. And when I think of stars, the overall skill of a player is what I envision those stars represent.

I bet if you used those metrics in that post, Darrach would have more stars than Dani. Or at least it would make more sense to me for Darrach to have more stars. Maybe Dani has more potential stars because he's younger and the scouts think he has more long term potential? But overall, it doesn't make sense for Dani to be higher (in stars) than Darrach.

I'm curious if changing those metrics would support that or not.

Ya'll thought I was joking about Darrach's BTT rating According to the "overall ratings" he's always been rated like a below average and yet he consistently plays at an above average level over his whole career (at a variety of positions too). You can see he's probably on the twilight end of things, but I bet he still has a few good seasons left. At his peak he was a 4+ WAR player which is all star level performance and now he's 2.5-3 WAR guy which is a solidly better than average player (and a superstar for a team like mine). Considering he was on 490k to $2Million/year salary for most of that time you can see hes' been a great value. For the 2049 season I'd give him a 55 OVR rating as a 3rd baseman, and a 65 BTT rating due to his other factors (versatility, leadership, and statistical consistency). Even though he's making more money now I didn't like the free agent options I had in the infield and I didn't like my situation in prospects for at least a few more seasons, so I still think he'll provide me good value at a 55 3B rating and 65 BTT rating on like $6mill/year with options each season.

 

[OU11]

Ok. Never mind than

All this changes is how the stars are presented, it does nothing for the actual ratings. They'll improve or lower as they always have. This just means if your rookie starts out slow and goes on a second half tear his current star rating will improve more than it did before since before it would have already been high to start

 

[Lloyd Carr]

I'm a big fan of Darrach. If I had to rate him, I'd consistently rate him a 3-3.5 star player.

 

[Gooksta]

Can't wait for a productive player with a good star ratings get picked up by a bad team and have his stars plummet.. lol

 

[osick87]

Can you do two different weights for current and potential?

 

[Travis7401]

I still think there is a problem with the way the overall ratings are calculated regardless, because they will not end up on a normal distribution as they probably should. Again this is because if you base ratings on a normal distribution then you cannot perform mathematical operations on them as if they were created with a uniform distribution (ie you cannot combine them with other ratings based on a separate normal distribution and then take the average). I would honestly have to see player ratings for the entire population to see how they were distributed before I had any comment on the usefulness of changing any of those settings, because I believe there is a fundamental mathematical error in the OOTP system for all the "overall" calculations.

I'll be using my own rating system anyway and I'll share it with anyone who wants, so I guess I don't really care what system you end up choosing... I'm just saying the underlying mathematical problem of combining normally distributed data is probably there no matter how you distribute the relative weights.

 

[OU11]

Using 40-20-25-15

Luong 3.5/3.5
Dani 1/3.5

Using 25s

Same

Using 25-30-30-15

Luong 4/4
Dani 1/3.5

Using 0-30-40-30

4/4
1/3.5

For a look at prospects, Dani Gut (my OF) is 1/5

 

[Karl Hungus]

@Travis7401

You should post this stuff over at the OOTP website. They'd probably pay attention.

 

[OU11]

Can you do two different weights for current and potential?

No, I think potential is just based on ratings no matter what since you can't really use past performance for potential

 

[OU11]

I still think there is a problem with the way the overall ratings are calculated regardless, because they will not end up on a normal distribution as they probably should. Again this is because if you base ratings on a normal distribution then you cannot perform mathematical operations on them as if they were created with a uniform distribution (ie you cannot combine them with other ratings based on a separate normal distribution and then take the average). I would honestly have to see player ratings for the entire population to see how they were distributed before I had any comment on the usefulness of changing any of those settings, because I believe there is a fundamental mathematical error in the OOTP system for all the "overall" calculations.

I'll be using my own rating system anyway and I'll share it with anyone who wants, so I guess I don't really care what system you end up choosing... I'm just saying the underlying mathematical problem of combining normally distributed data is probably there no matter how you distribute the relative weights.

Seriously if you want to get this deep, I'll make you a commissioner so you can mess with the settings. Just don't be a fag and look at the hard coded ratings to cheat.

 

[Travis7401]

@Travis7401

You should post this stuff over at the OOTP website. They'd probably pay attention.

I'll do so after I've done some more analysis.

 

[Travis7401]

Seriously if you want to get this deep, I'll make you a commissioner so you can mess with the settings. Just don't be a fag and look at the hard coded ratings to cheat.

DO NOT WANT

Anyway, this might be a better offseason discussion once I've got a data set to work with. I'll offer to help you evaluate different overall rating calcs at the end of the season to see if any of them provide useful information, then I can present the data with graphs and people can vote on how they'd like it displayed (and understand what the ratings mean). I just think the start of the season is probably the wrong time.

 

[OU11]

I'll do so after I've done some more analysis.

Really I'm thankful for this, I'm far too lazy to do this and this could expose some settings miscues I have in the league. For example if you could do analysis on how players age or how fast prospects develop I can tweak those settings. There are numerous ways this kind of data will improve the league going forward.

 

[OU11]

DO NOT WANT

Anyway, this might be a better offseason discussion once I've got a data set to work with. I'll offer to help you evaluate different overall rating calcs at the end of the season to see if any of them provide useful information, then I can present the data with graphs and people can vote on how they'd like it displayed (and understand what the ratings mean). I just think the start of the season is probably the wrong time.

Making you a commissioner doesn't mean you do any extra work, it just means you can evaluate different ratings systems compared to the stats put forth in your analysis. You can change the stars to overall, and input them into your data to look at which settings create the best overall dataset for the league.

 

[Orlando]

Would 40/20/25/15 be good? Two years ago is just a long time ago to put more weight into than the current year. Guys can fall off a cliff in a year which would make their overall skewed

I don't get putting much weight at all into the current year. It's such a small sample and the two prior years are a much better indicator.

 

[Travis7401]

I must say, hypothetically, that if this was a stealth Jihad mission and @kella put me up to infiltrate and destroy the WBL from the inside, then it is going swimmingly so far

1. Join the league
2. Convince them I'm at 80/80 immersion and have them make me a commissioner
3. ?
4. Destroy the WBL, profit.

 

[Gooksta]

I don't get putting much weight at all into the current year. It's such a small sample and the two prior years are a much better indicator.

Current year should be the Heaviest weight imo..what they are now is most important.. contract/trade situations

 

[OU11]

I don't get putting much weight at all into the current year. It's such a small sample and the two prior years are a much better indicator.

So that if a guy is hitting .200 throughout this year (which is all I'm concerned about honestly) he isn't rated a 5* because he hit .330 the previous two. That's just my thinking behind it, two years ago means nothing to me either. A lot can change in two seasons and I'd rather his current overall be based more on current events.

 

[Lloyd Carr]

I don't get putting much weight at all into the current year. It's such a small sample and the two prior years are a much better indicator.

Yeah, 40/10/25/25 might be a better ratio.

 

[OU11]

I must say, hypothetically, that if this was a stealth Jihad mission and @kella put me up to infiltrate and destroy the WBL from the inside, then it is going swimmingly so far

1. Join the league
2. Convince them I'm at 80/80 immersion and have them make me a commissioner
3. ?
4. Destroy the WBL, profit.

There isn't a way you can destroy the WBL as a commissioner. You'd have to upload long enough for you to know there are no backups and slowly sabotage it for weeks without people knowing until it was too far gone

When I hand it over to osick you can work through him though

 

[Travis7401]

I vote to leave it as is for now because we are already in season and I'll run some experiments to test how the distributions look based on various options and then we can discuss the pros and cons with a more detailed understanding of what each potential option means.

 

[Orlando]

Seriously if you want to get this deep, I'll make you a commissioner so you can mess with the settings. Just don't be a fag and look at the hard coded ratings to cheat.

You know those numbers change right? Unless you changed something for our league, when I ran player progression sims with a single player MLB save those numbers (both current and potential) changed.

 

[OU11]

I vote to leave it as is for now because we are already in season and I'll run some experiments to test how the distributions look based on various options and then we can discuss the pros and cons with a more detailed understanding of what each potential option means.

That's not how we do things here, we just change options for fun and face the consequences later.

 

[OU11]

You know those numbers change right? Unless you changed something for our league, when I ran player progression sims with a single player MLB save those numbers (both current and potential) changed.

Yeah, so if I just decided to look into all of the players real ratings it wouldn't be cheating because they could change? Nah b, if the league doeesn't have access, I don't have access when it comes to anything player related.

Besides nicknames of course.

 

[Orlando]

So that if a guy is hitting .200 throughout this year (which is all I'm concerned about honestly) he isn't rated a 5* because he hit .330 the previous two. That's just my thinking behind it, two years ago means nothing to me either. A lot can change in two seasons and I'd rather his current overall be based more on current events.

It just kills someone ratings for being in the midst of a cold streak.