My original league had
Pitching: Wins (W), Saves (SV), Outs (OUT), Strikeouts (K), Earned Run Average (ERA), Strikeouts per Nine Innings (K/9); 6 total
Pitching stats adjusted so 3 stats for starters (W, OUT, K), 3 stats for relievers (SV, K/9, ERA)
One of managers posted this in response
These are all categories that apply to RP slot with same or increased weight or benefit to RPs versus SP due to more K, lower ERA, and more K/9 innings pitched for the majority of players with RP status or those with mixed SP & RP eligible.
With an innings limit an out for a RP is same as that for a SP, A win and a SV are of equal value in a Roto rank but few or no 30+ wins pitchers and multiple 30+ saves pitchers.
In short, suspect I can skew things here by going after RP versus SP. Categories need more balance between SP-only Stats and RP or SP common stats which favor RP-eligible players.
This was my response:
Mike raised a concern about the pitching categories. In short, I've thought about it, and am considering the switch to QS, S, ERA, Outs-IPP, Strikeouts, and BB/9. This describes my thought process, a review of what I had, and a consideration to the suggestion given to me.
The controversial categories are here:
>Saves (SV), Outs (OUT), Strikeouts (K), Earned Run Average (ERA), Strikeouts per Nine Innings (K/9)
Wins, Saves, Strikeouts, Outs are counting statistics. The accumulate over the season.
ERA and K/9 are rate statistics.
Inherently starters would win out on counting statistics, relievers would win out rate statistics. Saves can only be won by relievers, so is 3 starters-3 reliever categories.
In detail, each category:
[*]Outs are essentially innings pitched, as far as I understand. If a pitcher gets 3 outs, he's notched for an inning pitched. Starters toss for 6 innings so, they have an inherent advantage. If outs are not anyway related IPP, then I will change to IPP.
[*]Wins can be won by starters and relievers. Quality starts can be considered as equal in weight but exclusive to starters which is a good counterpart to saves.
[*]Saves can only be won by relievers.
[*]Strikeouts and K/9 are essentially the same, a display of degeneracy (the linear algebra definition). I contemplated using K/BB instead of K/9, but there's no rigid correlation to K/BB to the quality of the start; think Scott Baker. Strikeouts favor starters because they're on the field longer**, K/9 favors relievers because relievers are chosen for their strikeouts in limited innings -- there's no groundball or flyball relievers except for Mo's cutter.
**The leading relievers in strikeouts last year were Aroldis Chapman (122) and Craig Kimbrel (116). There are 75 starters from last year who had higher in strikeouts.
[*]ERA is the rate statistic applicable to both relievers and starters.
If I change the pitching scoring categories I could imagine this:
QS - exclusive to starters,
S - exlusive to relievers,
ERA - the absolute pitching stat;
outs/IPP - another fundamental statistic; the pitcher's goal is to get as many outs as possible
K/BB - relievers also dominate K/BB like K/9, due to the same strikeout statistic, but not overwhelmingly. I might replace K/9 with this or BB/9. The correlation of K/BB to ERA is strong, leading to the basis behind FIP, but correlation does not always pan out.
strikeouts - which characterize many starters and relievers
If I had a team of starters only, they win: QS - Outs/IPP - Strikeouts.
If I had a team of relievers only, they win: S - K/BB - ERA
This fits the model of strikeout starters and relievers, yes. So what happens to the groundball or flyball pitcher who gets their outs by doing everything but strikeouts?
He wins QS and Outs/IPP depending on his ability. He gains a small advantage of K/BB since less strikeouts means less wild pitches so less BB. But K has the equal weight to BB (1 KB / 1 BB ) so the influence is negligible in the end. BB/9 would then be more "fair" than K/BB, for non-strikeout starters.
BB/9 is neutral to both relievers to starters. A small sample size doesn't shift bias towards each, so the final hypothetical categories:
If I had a team of starters only, they win: QS - Outs/IPP - Strikeouts
If I had a team of relievers only, they win: S - ERA
The final categories are
QS - S - ERA - Outs - Strikeouts - BB/9
So how well do these categories reflect on reality? Starters usually go 200 Innings. Relievers usually go up to 60 innings. So the innings were restricted to 1 starter, 1 reliever, where would be in total of 260 innings. Relievers would go 60 of the 260 innings or (23%). Starters would go 200 of the 260 innings or (77%). Now the analysis of our 6 categories [ QS – S – ERA – Outs – K – BB/9 ] .
Assuming the person has at least 4 starters - per restrictions, then the ERA advantage for relievers diminishes as the small sample size diminishes. ERA becomes like BB/9 or neutral to sample size quality. That leaves 3 inherent "starter" categories (3/4 = 75%), and 1 inherent "reliever" categories (1/4 = 25%).
Correct me if I'm wrong.