Individual Pitch vs Stuff/Mov/Con ratings?

[Matt Arnold]

Originally Posted by Anyone

To some degree (though the effect shouldn't be huge) great defense will lower a pitchers K/9. Any outs you get because a fielder made the play are chances you don't get to strike the batter out, so would lead to a better ERA but a worse K/9.

A really, really bad defense could inflate K/9's a lot actually, by keeping the inning going until the pirtcher struck guys out. But no one, including the AI, would play a defense that awful.

Yeah, generally speaking, K% (K/PA) should be constant regardless of the D behind you, but K/9 could vary depending on how many outs you record. I mean, take the extreme of a defense so bad that the pitcher has a 1.000 BABIP, then literally every out recorded is by strikeout, so the pitcher would by definition have a 27 K/9 whether he strikes out 3 batters or 300 batters.

[KBLover]

Taking a look at them through the lens of K%:

Casteles
2032: 244 K, 751 BF = 32.4 K%
2033: 260 K, 852 BF = 30.5 K%

Tatum
2032: 215 K, 587 BF = 36.6 K%
2033: 92 K, 235 BF = 39.1 K% (injury shortened season)

I guess it's just the way Casteles is and/or the defense impacts.

Not that I'm complaining about 30-32% K rate, at least that would put him in the top 5-10 pitchers in the game last year using MLB's 2019 numbers.

[CBeisbol]

Originally Posted by Matt Arnold

Yeah, generally speaking, K% (K/PA) should be constant regardless of the D behind you, but K/9 could vary depending on how many outs you record. I mean, take the extreme of a defense so bad that the pitcher has a 1.000 BABIP, then literally every out recorded is by strikeout, so the pitcher would by definition have a 27 K/9 whether he strikes out 3 batters or 300 batters.

So why does OOTP, in OOTP19 at least, use K/9, BB/9, etc, instead of the far superior K% or BB%?

Per inning stats for pitchers are obviously misunderstood - understandably since they are misleading.

[CBeisbol]

Originally Posted by KBLover

Taking a look at them through the lens of K%:

Casteles
2032: 244 K, 751 BF = 32.4 K%
2033: 260 K, 852 BF = 30.5 K%

Tatum
2032: 215 K, 587 BF = 36.6 K%
2033: 92 K, 235 BF = 39.1 K% (injury shortened season)

I guess it's just the way Casteles is and/or the defense impacts.

Not that I'm complaining about 30-32% K rate, at least that would put him in the top 5-10 pitchers in the game last year using MLB's 2019 numbers.

Compare the catchers.

Catchers impact K and BB rates.

See this thread, starting with post 11

[whifflebird]

It looks like the AI couldn't predict Hinch pulling out Grienke with only 80 pitches and having given up just two hits.

[Argonaut]

It looks like Casteles has the better overall defense, but yes, I'd be curious whether or not Tatum has the better catcher.

I've been similarly disappointed with my Nolan Ryan's K/9 rates, but it's happening because my defense is too good.

The best way to increase strikeouts is to have a bad defense and a great catcher.

[KBLover]

Originally Posted by CBeisbol

So why does OOTP, in OOTP19 at least, use K/9, BB/9, etc, instead of the far superior K% or BB%?

Per inning stats for pitchers are obviously misunderstood - understandably since they are misleading.

Yeah I noticed that when I was going to look up the two pitchers by K% and was like "huh, it's not there".

Batters have it but not pitchers. At the very least have both - maybe the /9 by default but the rates can be customized in.

[KBLover]

Originally Posted by CBeisbol

Compare the catchers.

Catchers impact K and BB rates.

See this thread, starting with post 11

(Editor ratings shown)

Casteles has a platoon:
Starter: Reece McGuire: 123 catcher ability
Bench (every second game): Jean-Pierre Sabatier: 99 catcher ability

So average, assuming they get the 50/50 split is 111 catcher ability.

Tatum has Ruiz with a 109 catcher ability

On the player profile, both McGurie and Ruiz show with 55 catcher ability. Sabatier has a 50.

[Argonaut]

Interesting. If you open Casteles and Tatum up in the editor, what kind of expected stats do their ratings spit out? You can check expected K per batter faced there.

If their expected stats have about the same K/PA... maybe try running the simulation module and checking those results. Could be some logic with the defenses that I can't ponder right now.

[KBLover]

Casteles:
956 BF, 409 K, 42.8 K%

Tatum:
982 BF, 409 K, 41.6 K%


Neither pitcher has been at these levels and it doesn't get more "modern, Major League environment" than using 2018's and now 2019's MLB stats.

I ran the Sim model with these two teams pit against each other.

Casteles: 246 K, 657 BF, 37.4 K%
Tatum: 254 K, 818 BF, 31.1 K%

Casteles' team had better DER (.726 vs .709). This didn't surprise me too much, the Foxes (Casteles' team) almost always strikeout fewer times than average. Tatum was still able to get them via K 31%, which is impressive.


Then I ran both of them against the Phillies, the NL's projected best offense.

Casteles: 257 K, 710 BF, 36.2 K%
Tatum: 261 K, 716 BF, 36.4 K%

Casteles' team had (slightly) better DER vs Phillies (.738 vs .733)

[Argonaut]

Interesting, thanks for those results. They look pretty in-line to me?

So I would wager what you're seeing from the game results is just some variance that to the human eye looks like a pattern.

Humans are very good at seeing patterns, it's an evolutionary thing.

[CBeisbol]

Yeah, when they played against the same offense they got the same number of K's.

One final test would be to copy one of them, put them on the same team and have them face the same team. Control as many variables as possible.

Same defense
Same catcher
Same offense
Same stadium
Same league environment

Based on the last post I'd expect same strikeouts (over a large enough sample - one season leaves a lot of room for variation)

[KBLover]

Originally Posted by Argonaut

Interesting, thanks for those results. They look pretty in-line to me?

So I would wager what you're seeing from the game results is just some variance that to the human eye looks like a pattern.

Humans are very good at seeing patterns, it's an evolutionary thing.

Interesting - but why was it persistent over multiple seasons? I guess just Casteles "low rolled" each season. Will be interesting to see how this season goes. So far, he's up there in K% (just hit the draft pool reveal).

Because if he was putting up 36%, I probably wouldn't even make the post lol.

Originally Posted by CBeisbol

Yeah, when they played against the same offense they got the same number of K's.

One final test would be to copy one of them, put them on the same team and have them face the same team. Control as many variables as possible.

Same defense
Same catcher
Same offense
Same stadium
Same league environment

Based on the last post I'd expect same strikeouts (over a large enough sample - one season leaves a lot of room for variation)

Since a thought was that defense behind Casteles was a possible cause for his lower total, I put Tatum on Casteles' team.

Once again, sim model vs Phillies. I ran it three times.

"Season 1"
Casteles: 256 K, 726 BF, 35.3 K%
Tatum: 259 K, 674 BF, 38.4 K%

"Season 2"
Casteles: 249 K, 703 BF, 35.4 K%
Tatum: 221 K, 705 BF, 32.2 K%

"Season 3"
Casteles: 261 K, 671 BF, 38.9 K%
Tatum: 249 K, 709 BF, 35.1 K%

Totals:
Casteles: 36.5 K%
Tatum: 34.9 K%

Still at a loss why the sim model is putting up 35% or higher consistently for Casteles while during the seasons, he was barely squeezing out 30% consistently.

[Argonaut]

I don't know what division you put Montreal in. Maybe that division has batters with higher Avoid K? Or maybe you put Montreal in the AL and he pitches against DH's while the STL pitcher nets extra K's against pitchers?

[KBLover]

Originally Posted by Argonaut

I don't know what division you put Montreal in. Maybe that division has batters with higher Avoid K? Or maybe you put Montreal in the AL and he pitches against DH's while the STL pitcher nets extra K's against pitchers?

Both leagues have the DH now and that's been the case since both have been in the league.

It could be the divisions - NL East (where Montreal is) is a bit better than NL average in Avoiding Ks. Montreal's park would probably be considered a slight pitcher's park (HR down, strongly down vs RHB but doubles/triples significantly higher than average).

Oh well - it is what it is. So far this season (at the draft), Casteles is at 39% and Tatum 37%, so maybe it's evening out or something.