Look at the exp win %. That will tell you all you need to know! LMAO!!

Nice try Bob. You were knocked out in the first round of the facts discussion. Keep trying to get up, the result will be the same. Please spare all the innocent on this thread from another weak attempt by you to mock someone who disagrees with you and provided facts to support their opinion. Which you do not.

You are like the guy in the classic Monte Python movie who kept shouting "flesh wound" as his opponent kept slicing off limbs....hahahahahah.

In order of importance:
OAV (and it's + and # variants) is most important. It's one of three main elements on the decision tree, and the most influential one.
Ballpark - the impact a park has is more significant that most realize.
Season, depending on ballpark and defense, I either won't draft a pitcher pre-1980 or post-1936 unless a theme requires it or there's a one-off pitcher that is far enough outside the talent scope of the rest of the availables.
BB/9 (and it's + and # variants) is another I look at, though it's one of three main elements on the decision tree, it's value is mostly dependent on ballpark and defense
HR/9 (and it's + and # variants) is another I look at, though it's one of three main elements on the decision tree, it's value is mostly dependent on ballpark, defense, and season of pitchers chosen

In most leagues I completely ignore HR/9. The Modern vs Deadball TWISL cemented that for similarly priced pitchers, regardless of HR and K, you get similar value out of them, and I find it easier to control the weaknesses of modern pitchers with HR and defense than I do deadball pitchers, so I find more value from 1980+ pitchers than from the deadball ones despite low OAV/WHIP/BB/HR.

These are the right metrics, but their order of priority must be weighted based on your park (e.g. bb/9# more relevant in Petco, oav# and hr/9# more relevant in Coors). Erc# and $/ip is your first filter, and build for your park and defense from there. Erc# is kind of a proxy for real-life FIP (without the k component), and the other metrics are constituents of that variable, so you can play with how you weight them. If you're already paying for defensive range, or in Petco, it's redundant and inefficient to pay up for oav#. Your $ is better spent elsewhere, as the $ will not be discounted.

Lastly, of course you must be aware of ip/g. What I have played with most lately is guys with odd ip/g in tandems. They are often at a $ discount for their erc# but you need to manipulate their availability. Tandems are great for this. I also like using a better, low IP/g guy at the front of a tandem to guarantee he faces the heart of a lineup (e.g. a 30 pitch tandem a should get you through the top of a lineup two times, while your weaker tandem b will turn them over just once). That's a way to pay higher $/IP for a guy with fewer total IP, which saves $ overall. It still maximizes where your good $ are deployed.

My questions - how is fatigue or performance affected when using a SP as RP, or vice versa? Does this also apply to tandems? If I use a SP as a tandem B, or even as a regular reliever, is their performance getting dinged, and how so? Even if you have to factor erc# by 0.9, that changes all of the above math.

In regards to IP/G… I generally don’t filter for it or look at it when drafting, but I pay particular attention to it when setting up my rotations and pitch counts. I tend to draft for a specific inning count (1000-1400) depending on park and salary cap and not worry about how I get to that total, just look at taking the best players in my budget. Sometimes that means I have thirteen 90 inning pitchers, sometimes I have one 265 IP guy and a bunch between 80-130 with a few in the 25-50 range. I then use IP/G to determine roles and use after I’ve drafted. That said, there are a large number of bargain pitchers in the <5 IP/G range and in the 90-160 IP range. Modern ~1 IP/G guys are also usually underpriced, especially in regards to how far you can push them in terms of fatigue. Their PC prices in the high K rate, normalization assures they won’t use it, so even a 0.80 IP/G can usually go on a 15 PC without issue and if has 70+ IP is easily good for 80+ appearances. (Thirteen ~90 IP pitchers with ~1 IP/G will essentially have each pitcher pitch 1 inning per game and you only usually have issues in/with extra inning games).

I started testing SP in relief, though I don’t recall if I finished those tests. It was clear that SPs seem to have a built in fatigue wall of around 100 pitches. If I recall correctly, the testing was going to next verify if that was a hard number or if it was a % based on IP/G (all the pitchers in the first test had similar high IP/G). I don’t recall if I ever did the second test with lower IP/G SPs in relief to see if 100 pitches was still the fatigue rule of thumb. This seemed to apply to the second pitcher in a tandem just as much as an SP in an RP role. I’ve seen no issues with RPs as SPs other than making sure you’re not exceeding their IP/G.