New Proposed Stat: Win Or Quality Start

Although it’s not quite an age-old debate, we in the fantasy baseball community like to argue the relative merits of using the win or quality start (henceforth QS) for measuring pitcher performance. Obviously, in a general baseball sense, we mostly agree that neither stat is important. However, for fantasy purposes, a count of good starts does seem a useful category. And we only have two ways to do it – either use wins or QS. I’m here to propose a third alternative.

For the record, if forced to pick a side in a straight battle, I’ll choose wins every time. I really don’t view QS as a viable fantasy category except in the shallowest of leagues. They’re exceptionally predictable – mostly because the supply of pitchers who regularly finish the sixth inning is tiny. Proponents of the QS call that a feature. I think it’s a bug. The pool of pitchers who can supply a rosterable total of QS is typically between 40 and 50 players – depending on how you define “rosterable.” Our friends (competitors?) at Baseball Prospectus found that the QS is dying.

From a narrow view, wins are even harder to buy in bulk. More pitchers reach 10 QS than 10 wins. If you consider both stats on a rate basis – that is QS/IP and W/IP – that calculus changes dramatically. There are still just a few dozen pitchers who earn QS in bulk. Some of them are guys like Jordan Zimmermann – i.e. unrosterable for other reasons. When shiny prospects receive their call up, they aren’t very likely to pitch six innings. They do have a shot at a win.

Additionally, wins do not deprive relievers of a category. This has massive implications for player value, and it leaves open a wide range of strategic options for roster management. Since most non-elite starters and quality relievers win games at roughly the same rate by W/IP, owners can literally turn to a pool of hundreds to address their needs. This is especially important in head-to-head formats. We know there will always be 2,430 wins in a season (less rainouts).

An Alternative

Enemies of the win point to the fickle nature of the stat. A pitcher could cough up seven runs in five innings and record the victory. Relievers vulture wins all the time – that is, they give up the lead and then their offense immediately responds in the next half inning. The worst is when a pitcher throws eight innings of one-run ball en route to a no decision. Some pitchers on bad teams (ahem, Chris Archer), frequently win fewer than 10 games despite 15 win ability. You always know the exact parameters of the QS. Offense isn’t a factor.

In the intro paragraph, I identified what we’re trying to capture – a proxy for a count of good starts. Wins don’t quite do it. Run support and reliever meltdowns scramble the calculus. QS work in theory but not in practice due to their limited supply. They also exclude relievers. QS leagues are strategically hobbled.

My proposal is to combine the stats into WorQS (rhymes with “corks”). Or if you prefer WandorQS (“wan dorks” or “wand orcs”). This is pretty simple. If a pitcher records a standard win, he earns one WorQS. If he tosses a QS, he earns one WorQS. And here’s the wrinkle – if he gets a win AND a QS, he still earns only one WorQS. The point of the stat is to eat away at some of the downsides for both stats. In the case of wins, this eliminates the need for run or reliever support. It also fixes the supply-side problem for the QS.

Are there issues? Sure. Mostly, undeserved wins will still be awarded to bad starts and vulture relievers. I don’t see that as a big problem. After all, those pitchers are still tagged with lethal rate stats for those outings.