Switch to a better RNG for ABE

Let us know what you'd like to see!
ratman44
Posts: 1481
Joined: Sat Sep 30, 2017 9:35 am

Re: The most talked about topic on the whole forum

Post by ratman44 »

My only complaint is when an A overall pitcher backed by a stacked lineup which faces a B overall pitcher with a rebuilding lineup where no one from that lineup would even be on the better team. The B overall pitcher should not win the game. But it seems to happen more than I care to admit.

I know that is the randomness, but considering I have guys in my minors better than the guys I am playing it makes the whole thing humorous.

Perfect example...

http://simdynasty.com/boxscore.jsp?boxs ... &cityid=13

I know it is just 1 game, but it is really frustrating to watch.
User avatar
ballmark
Posts: 2774
Joined: Sat Sep 30, 2017 8:33 am
Location: Michigan
Contact:

Re: The most talked about topic on the whole forum

Post by ballmark »

Have you met Bundy...?
User avatar
Vytron
Posts: 555
Joined: Sat Mar 30, 2019 4:27 am

Re: The most talked about topic on the whole forum

Post by Vytron »

ratman44 wrote: Wed Mar 23, 2022 11:33 pm But it seems to happen more than I care to admit.
SPECIALLY if the better pitcher is an A-/A-, they regularly get beaten by B+/B+ ones. Here's an example:

http://www.simdynasty.com/player.jsp?mo ... d=13917138

The league is reset tomorrow so I'll take screen shoots...

Here's the pitcher's stats:

Image

How'd you expect her to perform? Here's her improvement chart:

Image

It's not much, but she's improving a bit. And now, here's reality:

Image

On a "I'm gonna suck as a pitcher" pattern, and a "I'm gonna suck more and more as a pitcher" pattern. And I know, because I babysat her and moved her from minors to majors and back so she would both pitch and be on the minors for improvement chances, which worked excellently to get converted ICs at her age, but it's all useless when randomness overcomes them like this. Better to trade her for some B+/B+ pitcher that isn't on that pattern? That shouldn't work...

Oh, and let's not forget...

http://www.simdynasty.com/player.jsp?mo ... d=13938513

Image

B+ Velocity A Control pitchers are supposed to be good! But

Image

The jump of pattern is noticeable between seasons, as if she was lobotomized.

If this was all it'd be bad luck and fine, but once on these patterns you can predict they'll continue to suck for the rest of the season, at the point ABE switches patterns (but they can be put on an even worse one!). Players are consistently overperforming or underperforming, not a just one game thing, but for a half-season.

At least this isn't the most talked about topic anymore (if you exclude this thread), it has been taken over by Congratulations threads :mrgreen:
SquidJordan wrote: Fri May 13, 2022 12:09 am all you can do is all you can do.
Vytron wrote: Sun May 22, 2022 10:01 pm Disclaimer: I don't know what I'm talking about
ratman44
Posts: 1481
Joined: Sat Sep 30, 2017 9:35 am

Re: The most talked about topic on the whole forum

Post by ratman44 »

Vytron wrote: Thu Mar 24, 2022 8:49 am
ratman44 wrote: Wed Mar 23, 2022 11:33 pm But it seems to happen more than I care to admit.
SPECIALLY if the better pitcher is an A-/A-, they regularly get beaten by B+/B+ ones. Here's an example:


If this was all it'd be bad luck and fine, but once on these patterns you can predict they'll continue to suck for the rest of the season, at the point ABE switches patterns (but they can be put on an even worse one!). Players are consistently overperforming or underperforming, not a just one game thing, but for a half-season.
So 2 things B+/B+ is really not that much different than A-/A- so I am completely ok with that. It’s possible you have a 75/75 pitcher beating a 76/76 pitcher. They are basically the same pitcher.


Then you posted 2 pitchers having 33 innings of bad pitching saying it doesn’t reverse. All you need is one bad outing to ruin your ERA in 33 innings.

I had my ace lost his first 3 games one season giving up 15 runs in 9 innings. With your logic I should have abandoned him. But he ended the season 20-7 with a 2.60 ERA.

At the same time here is a thread about my 3B who was hitting.186 60 games in. But he ended the season .278 so it evened out.

viewtopic.php?f=68&t=34783
User avatar
Vytron
Posts: 555
Joined: Sat Mar 30, 2019 4:27 am

Re: The most talked about topic on the whole forum

Post by Vytron »

ratman44 wrote: Thu Mar 24, 2022 2:18 pmSo 2 things B+/B+ is really not that much different than A-/A- so I am completely ok with that.
Not when A-/A- pitchers are consistently worse than B+/A-. A-/B+ and B+/B+ ones. And I'm not talking about face to face, but the stats they end up at the end of the season (or, er, half-season, this is trial leagues). Is this some esoteric knowedge? Because I've seen people making weird trades and draft orders for players that ABE is going to like, as if they figured out something (but I refuse to point fingers, it's like some kind of subgame where luck can be predicted.)
ratman44 wrote: Thu Mar 24, 2022 2:18 pm All you need is one bad outing to ruin your ERA in 33 innings.
I'm talking about WHIP, not ERA. Heck, on the ERA side ABE favors me more than not, with terrible pitchers coming in and winning my games or saving them. What happens mostly is that I have a pitcher with terrible ERA but much more wins than loses, I like that.

About the other thing, yeah, I mentioned 40% of players are following these patterns, didn't I? So you can pile up counter-examples and more examples than me, I claim with a working RNG no predictions could be made at all because all outcomes would be independent, as they're programmed, but they're not (which works as it keeps people engaged trying to figure the sim our, but there's always people quitting because they don't know what's going on.)
SquidJordan wrote: Fri May 13, 2022 12:09 am all you can do is all you can do.
Vytron wrote: Sun May 22, 2022 10:01 pm Disclaimer: I don't know what I'm talking about
ratman44
Posts: 1481
Joined: Sat Sep 30, 2017 9:35 am

Re: The most talked about topic on the whole forum

Post by ratman44 »

Vytron wrote: Thu Mar 24, 2022 3:14 pm
Not when A-/A- pitchers are consistently worse than B+/A-. A-/B+ and B+/B+ ones. And I'm not talking about face to face, but the stats they end up at the end of the season (or, er, half-season, this is trial leagues).
But again I still wouldn’t see that as an issue as the difference in all the players you mentioned is statistically insignificant.

Unless you know the guy is 83/83 and the B+ is 68/68 your difference in level is minimal in most cases. A 76/76 A-/A- is actually a worse pitcher than a 74/83 B+/A-.

None of those guys are aces, they have inconsistent moments. The people I play with find A-/A- guys as back end rotation guys. From my historical data my spreadsheets don’t support your findings, but I only keep data on full seasons.

My issue is when my A overall A-/A pitcher gets spanked by an inferior lineup while an A+/C+ Pitcher shuts out my lineup.

But if people are figuring the sim out on exact players of inferior quality that’s pretty good stuff on their part. I know every so often power pitchers just go Ape, but I can never predict when.
User avatar
Vytron
Posts: 555
Joined: Sat Mar 30, 2019 4:27 am

Re: The most talked about topic on the whole forum

Post by Vytron »

ratman44 wrote: Thu Mar 24, 2022 7:40 pm None of those guys are aces, they have inconsistent moments.
The most inconsistent I've ever seen, and I suspect it's an RNG issue because I observe the same in A- Power A- Contact hitters (which can even be outmatched by B+ Power / B Contact hitters!)
ratman44 wrote: Thu Mar 24, 2022 7:40 pmFrom my historical data my spreadsheets don’t support your findings, but I only keep data on full seasons.
Are they supported if you split the data in half seasons? Do you see players overperforming in one half and underperforming in the other showcasing the "switch pattern" I mention?
ratman44 wrote: Thu Mar 24, 2022 7:40 pmMy issue is when my A overall A-/A pitcher gets spanked by an inferior lineup while an A+/C+ Pitcher shuts out my lineup.
The question is if your opponent knew about this and was able to predict it, drafting the A+/C+ higher or even trading a much better pitcher for the A+/C+ one, because they know some inner workings from ABE we don't.
SquidJordan wrote: Fri May 13, 2022 12:09 am all you can do is all you can do.
Vytron wrote: Sun May 22, 2022 10:01 pm Disclaimer: I don't know what I'm talking about
ratman44
Posts: 1481
Joined: Sat Sep 30, 2017 9:35 am

Re: The most talked about topic on the whole forum

Post by ratman44 »

I only download full year data, but the data you talk about mimics real life.

There are multiple historical players that are hot and cold. So someone will hit .410 in April/May but only .280 the rest of the way.

No there is nothing special about A+/C+ pitchers the dude won one game today, but by the end of the season his stats will regress back to terribleness. My team is just slumping to start the season, World Series hangover. It’s just annoying. But to be honest doesn’t everyone give the defending champs their best shot?
User avatar
Vytron
Posts: 555
Joined: Sat Mar 30, 2019 4:27 am

Re: The most talked about topic on the whole forum

Post by Vytron »

ratman44 wrote: Thu Mar 24, 2022 11:11 pmThere are multiple historical players that are hot and cold.
Yes, yes, this looks more like a feature than a bug, baseball simulators like History Maker Baseball have mechanics to mimic this by rolling dice to set who's slumping or inspired. Admin has said nothing of the sort has been programmed here, though, so that's weird.
SquidJordan wrote: Fri May 13, 2022 12:09 am all you can do is all you can do.
Vytron wrote: Sun May 22, 2022 10:01 pm Disclaimer: I don't know what I'm talking about
eragon
Posts: 80
Joined: Sun Oct 22, 2017 10:02 pm

Re: The most talked about topic on the whole forum

Post by eragon »

Vytron wrote: Wed Mar 09, 2022 10:44 pm
bygmester wrote: Wed Mar 09, 2022 8:24 am If what you say is accurate it shouldn't be difficult to find some real time examples to post here on the forum.
Yeah, the reason I haven't been posting them is because anybody could say "with so few at bats anything can happen", and they'd be right. That's why pseudorandomness is so widely used, it's incredibly hard to prove its biases.

I'm playing in the maximum 10 trial leagues possible, so I'm playing 10 half seasons per month, but anyway:

Overperforming:

http://www.simdynasty.com/player.jsp?mo ... d=13934834
http://www.simdynasty.com/player.jsp?mo ... d=13934661
http://www.simdynasty.com/player.jsp?mo ... d=13917243
http://www.simdynasty.com/player.jsp?mo ... d=13914492
http://www.simdynasty.com/player.jsp?mo ... d=13914490
http://www.simdynasty.com/player.jsp?mo ... d=13914715
http://www.simdynasty.com/player.jsp?mo ... d=13914499
http://www.simdynasty.com/player.jsp?mo ... d=13916473
http://www.simdynasty.com/player.jsp?mo ... d=13850449
http://www.simdynasty.com/player.jsp?mo ... d=13850199

Underperforming:

http://www.simdynasty.com/player.jsp?mo ... d=13934330
http://www.simdynasty.com/player.jsp?mo ... d=13934321
http://www.simdynasty.com/player.jsp?mo ... d=13917044
http://www.simdynasty.com/player.jsp?mo ... d=13917210
http://www.simdynasty.com/player.jsp?mo ... d=13914873
http://www.simdynasty.com/player.jsp?mo ... d=13914839
http://www.simdynasty.com/player.jsp?mo ... d=13916017
http://www.simdynasty.com/player.jsp?mo ... d=13850136

...

These are from half my leagues...

Expected behavior: Since future results will be random, you can't use the players' data to predict anything.

What actually happens: Most of the overperforming players will continue to overperform for the rest of the season, and most of the players underperforming will continue to underperform. A non-random trending pattern.

Basically the players on the bottom list should have the stats of the players on the top list, and viceversa, but randomness is overpowering letter grades.

Another claim I have no evidence for: Converted ICs and name changes put the players in a different pattern. It happened recently, I had most of my players underperforming in a league, hopelessly stuck in last place, so I returned all of them to their original names and abandoned the league. Their patterns changed and the team ended in third place a win away from the playoffs.

I'm not complaining (I have benefited from this), and maybe it's dumb from me to reveal this out (like a name change or sending a player to the minors to get a single converted IC can change their pattern to overperform), but I claim something's up.
Overperforming:
.671 OPS in 283 at bats
.665 OPS in 97 at bats
Underperforming:
.782 OPS in 499 at bats

You're looking for patterns that don't exist (which makes sense, that's how the human brain works, but this is not a conspiracy lolz)
Post Reply

Return to “Baseball Suggestions And Enhancements”