I've done this analysis before and the results always end up the same. But I wanted to repeat this just to see.

Results coming shortly.

The process.

Through 68 games, I downloaded every player's real life stats and sim stats. I used the pitcher's real life games started to games played ratio to differentiate SP vs RP. If the ratio > 60%, then the pitcher was labelled a "SP". If the ratio < 10%, then the pitcher is labelled a "RP". Anything in between, the pitcher is labeled "Mix".

I applied the sim usage (i.e., innings) to the real life stats to get an overall "projected" stat. Then I compared that to the sim stat, and compared the two stats. I did this for whip, oav, bb/9, hr/9 & era.

I will present the results of each of the four leagues below.

League #1 Results
The RPs are slightly worse than the SPs... I started thinking that maybe my thinking was biased.

		Proj	Proj	Proj	Proj	Proj
Type___	ProjIP___	WHIP___	OAV___	BB/9___	HR/9___	ERA
SP	10995	1.10	0.209	2.42	0.56	2.44
RP	3189	1.08	0.196	2.84	0.46	2.10
Mix	354	1.18	0.217	2.55	0.55	2.53
Total	14539	1.09	0.207	2.52	0.54	2.37
		Sim	Sim	Sim	Sim	Sim
Type	SimIP	WHIP	OAV	BB/9	HR/9	ERA
SP	10995	1.44	0.262	3.63	1.00	4.83
RP	3189	1.42	0.247	4.13	0.81	4.31
Mix	354	1.65	0.294	3.86	1.40	6.02
Total	14539	1.44	0.259	3.75	0.97	4.75
	Type	WHIP	OAV	BB/9	HR/9	ERA
	SP	131%	125%	150%	179%	198%
	RP	132%	126%	145%	177%	206%
	Mix	139%	136%	152%	253%	238%
	Total	132%	126%	149%	181%	200%

League #2 Results
Nope, not biased. Clearly, RPs perform worse relative to their stats than SPs.

		Proj	Proj	Proj	Proj	Proj
Type___	ProjIP___	WHIP___	OAV___	BB/9___	HR/9___	ERA
SP	10965	1.11	0.213	2.23	0.64	2.58
RP	3034	1.03	0.185	2.66	0.53	2.04
Mix	524	1.12	0.215	2.09	0.43	2.32
Total	14523	1.09	0.208	2.32	0.61	2.46
		Sim	Sim	Sim	Sim	Sim
Type	SimIP	WHIP	OAV	BB/9	HR/9	ERA
SP	10965	1.46	0.273	3.27	1.39	5.19
RP	3034	1.44	0.253	4.00	1.21	4.72
Mix	524	1.54	0.287	3.21	1.07	4.97
Total	14523	1.46	0.269	3.42	1.34	5.08
	Type	WHIP	OAV	BB/9	HR/9	ERA
	SP	132%	128%	146%	215%	201%
	RP	140%	136%	150%	229%	231%
	Mix	138%	133%	154%	247%	214%
	Total	134%	130%	148%	218%	207%

League #3 Results
.

		Proj	Proj	Proj	Proj	Proj
Type___	ProjIP___	WHIP___	OAV___	BB/9___	HR/9___	ERA
SP	10706	1.11	0.212	2.31	0.61	2.60
RP	3274	1.00	0.183	2.71	0.48	1.89
Mix	540	1.23	0.226	2.76	0.60	2.84
Total	14520	1.09	0.206	2.42	0.58	2.45
		Sim	Sim	Sim	Sim	Sim
Type	SimIP	WHIP	OAV	BB/9	HR/9	ERA
SP	10706	1.45	0.271	3.32	1.10	4.95
RP	3274	1.36	0.241	3.91	0.91	4.17
Mix	540	1.66	0.289	4.33	1.15	5.68
Total	14520	1.44	0.265	3.49	1.06	4.80
	Type	WHIP	OAV	BB/9	HR/9	ERA
	SP	131%	128%	144%	180%	191%
	RP	136%	132%	144%	190%	220%
	Mix	135%	128%	157%	191%	200%
	Total	132%	129%	144%	182%	196%

League #4 Results
.

		Proj	Proj	Proj	Proj	Proj
Type___	ProjIP___	WHIP___	OAV___	BB/9___	HR/9___	ERA
SP	10664	1.12	0.216	2.27	0.59	2.63
RP	3325	1.04	0.193	2.49	0.42	1.94
Mix	588	1.14	0.200	2.94	0.52	2.38
Total	14577	1.11	0.210	2.35	0.55	2.46
		Sim	Sim	Sim	Sim	Sim
Type	SimIP	WHIP	OAV	BB/9	HR/9	ERA
SP	10664	1.45	0.275	3.06	0.93	4.99
RP	3325	1.43	0.263	3.42	0.74	4.49
Mix	588	1.58	0.285	3.72	0.81	5.13
Total	14577	1.45	0.273	3.17	0.88	4.88
	Type	WHIP	OAV	BB/9	HR/9	ERA
	SP	129%	128%	135%	159%	190%
	RP	137%	136%	137%	176%	231%
	Mix	138%	143%	127%	155%	215%
	Total	131%	130%	135%	161%	198%

This is super interesting. I have so many questions and thoughts!

This doesn't surprise me, and I believe that the reason for this is in-game fatigue. A starting pitcher is going to have in-game fatigue applied to a much lower percentage of his plate appearances than a reliever does.

Let's say a starting pitcher with RL 6 IP/G is set to a PC of 90. He'll face about 25 hitters per game, and only the last PA will have in-game fatigue applied. That's 4% of his batters faced.

A reliever with RL 1.0 IP/G set to a PC of 15 might face 4 or 5 batters per game, and the last PA will have in-game fatigue applied, which is 20-25% of his batters faced.

I agree that in-game fatigue was one of my thoughts for possible explanations. But why would they have it applied if pitch counts are set correctly?

Posted by footballmm11 on 3/31/2025 2:04:00 PM (view original):
I agree that in-game fatigue was one of my thoughts for possible explanations. But why would they have it applied if pitch counts are set correctly?

Here's my theory. Let's say your RP has 1.0 RL IP/g and has his PC set to 15/15. He faces a few batters, after which he has thrown 13 pitches. He will likely be allowed to face the next batter. He throws 4 pitches to that batter, making it 17 total. It is my belief that, because the deciding pitch of that PA is the 17th, which is more than 15, in-game fatigue will be applied to that plate appearance.

I do not know whether this is how it actually works, but I believe that this is how it works, both because (at least to me) it makes logical sense and because experience (such as the data compiled by schwarze) suggests it.

As a quick test, I looked at three of my 2021 Padres relievers who are set to 15/15. They have combined for 87 appearances, and in 52 of those their pitch count exceeded 15. They have collectively faced 361 batters, meaning that, if my assumption is correct, 14.4% of their plate appearances pitched have had in-game fatigue applied.

I don't have similar stats for SPs with high pitch counts, because I use tandems an overwhelming majority of the time, but I'll bet their PA%'s with in-game fatigue tend to be even lower than the 4% I originally cited.

Seems plausible

Another factor that may have a minor impact is that some RPs get brought in late in the game (i.e., extra innings) that might start off less than 100%. I'm not going through every boxscore to see how often that happens. This is a league with no salary cap, so I would expect most teams to have enough innings on their roster.

I agree that RP in-game fatigue could slightly impact the numbers, although I purposely set most of my low IP/G relievers to 10/10. My RPs still mostly get killed (more frequently than the SPs).

Sadly, in a high salary cap environment, this often results in a stud closer like Rosenthal to be removed from the 9th inning before getting 3 outs... which means a lesser RP comes in and blows the game. It's more unlikely than likely for a RP to complete one inning on 10-12 pitches. (Although strangely enough, I've had a RP recently give up 6 hits and 6 runs in as few as 16 pitches).

Not until the "powers-to-be" change the program to eliminate or somehow modify pitch count as a basis for fatigue, nothing is going to change.