Some splits make no sense

joestros · 04-30-2016, 01:34 AM

I cannot for the life of me figure out how the game is figuring out these splits. For example, 2015 Jonathan Schoop was much better vs. righties (.301/.329/.563 vs. R, .232/.260/.313 vs L), yet in the game his LEFTY Contact, Gap, and Power ratings are all better. This makes no sense. He even had more ABs vs. R. Why would his lefty ratings be better when he was worse vs. lefties AND had less at bats vs. lefties?

Matt Arnold · 04-30-2016, 08:51 AM

You need a much larger sample to figure out splits. Players naturally are better against opposite-handed pitchers, so you can't overreact to a one year sample.

OakDragon · 04-30-2016, 11:40 AM

Two things:

1.: The splits are statistics, not tendencies, so they show how a player did, not what they're likely to do.

2.: At one time, splits were random, so any "tendency" you found was the luck of the draw. I don't know if this is still the case or not. Does OOTP now have L/R tendencies built in behind the scenes?

joestros · 04-30-2016, 11:58 AM

Quote:

Originally Posted by Matt Arnold

You need a much larger sample to figure out splits. Players naturally are better against opposite-handed pitchers, so you can't overreact to a one year sample.

But we are playing a single season replay so the ratings are supposed to be optimized for single season. I don't feel like it is correct to override a player's splits just because "players are naturally better against opposite-handed pitchers". OOTP simulation engine can't be taken seriously if it seemingly randomly assigns ratings to real players. I know many who play OOTP don't necessarily care about realism, but, if you are going to include options to replay a single season, and/or use real players, I feel like ratings should reflect how a player performed in real life.

Matt Arnold · 04-30-2016, 12:14 PM

For single season replays, there's certainly a better argument. But it can get out of line, and cause weird issues. For example, if you had Nolan Arenado batting against Justin Wilson, and you have Pedro Alvarez on your bench, would you have Alvarez pinch hit? Real life you would get laughed out of the league for doing that, but if you look at 2015 stats for all 3 players, you would realize that the Alvarez matchup is much more is actually very much in your favour.

joestros · 04-30-2016, 12:55 PM

For some league types, this is ok. There should be a setting that allows stats and ratings to be based purely on the last real season stats with maybe a small adjustment for low AB guys. For other leagues, maybe using a real-life 3-year average would work best.

Other options are to include some sort of playing time limit for low AB players so they could not be overused. Our league uses a self-imposed 150 AB limit to weed out low AB players (i.e. you cannot draft a player unless they had at least 150 ABs in real life). There are ways around this without messing with the ratings.

04-30-2016, 01:34 AM	#1
joestros Minors (Single A) Join Date: Jun 2002 Posts: 77	Some splits make no sense I cannot for the life of me figure out how the game is figuring out these splits. For example, 2015 Jonathan Schoop was much better vs. righties (.301/.329/.563 vs. R, .232/.260/.313 vs L), yet in the game his LEFTY Contact, Gap, and Power ratings are all better. This makes no sense. He even had more ABs vs. R. Why would his lefty ratings be better when he was worse vs. lefties AND had less at bats vs. lefties?

04-30-2016, 12:55 PM	#6
joestros Minors (Single A) Join Date: Jun 2002 Posts: 77	For some league types, this is ok. There should be a setting that allows stats and ratings to be based purely on the last real season stats with maybe a small adjustment for low AB guys. For other leagues, maybe using a real-life 3-year average would work best. Other options are to include some sort of playing time limit for low AB players so they could not be overused. Our league uses a self-imposed 150 AB limit to weed out low AB players (i.e. you cannot draft a player unless they had at least 150 ABs in real life). There are ways around this without messing with the ratings. Last edited by joestros; 04-30-2016 at 01:02 PM.

04-30-2016, 08:51 AM	#2
Matt Arnold OOTP Developer Join Date: Jun 2009 Location: Here and there Posts: 16,242	You need a much larger sample to figure out splits. Players naturally are better against opposite-handed pitchers, so you can't overreact to a one year sample.

04-30-2016, 11:40 AM	#3
OakDragon Hall Of Famer Join Date: Feb 2006 Location: San Jose, CA USA Posts: 3,494	Two things: 1.: The splits are statistics, not tendencies, so they show how a player did, not what they're likely to do. 2.: At one time, splits were random, so any "tendency" you found was the luck of the draw. I don't know if this is still the case or not. Does OOTP now have L/R tendencies built in behind the scenes?

04-30-2016, 12:14 PM	#5
Matt Arnold OOTP Developer Join Date: Jun 2009 Location: Here and there Posts: 16,242	For single season replays, there's certainly a better argument. But it can get out of line, and cause weird issues. For example, if you had Nolan Arenado batting against Justin Wilson, and you have Pedro Alvarez on your bench, would you have Alvarez pinch hit? Real life you would get laughed out of the league for doing that, but if you look at 2015 stats for all 3 players, you would realize that the Alvarez matchup is much more is actually very much in your favour.