How relevant are pitcher wins?


Traditionalists can keep fighting, but their battle for wins is a losing cause.

As pitching grows increasingly specialized -- there has been a lower rate of starts lasting at least 110 pitches this season (12 percent) than in any time during the past quarter-century (that encompassing all seasons for which Baseball-Reference.com has pitch-count data; in all likelihood it's an all-time low) -- the value of the win continues to tumble. Quality starts -- those of at least six innings pitched and no greater than three earned runs allowed -- have resulted in a win only 53.3 percent of the time, the second-lowest rate in almost a century (2011, 53.2 percent).

Baseball might be a game of statistical benchmarks, the 20-win plateau a popular such example, but in the modern game that specific accomplishment carries far less meaning; it's mostly the product of extremely good fortune.

Consider: In the 13 completed seasons since 2000, eight times the respective league's (American or National) pitching leader in Wins Above Replacement won 20 or more games … but nine times he won 16 games or fewer.

Using this season's returns to illustrate: Among the 11 pitchers on pace for 20 wins are the names Patrick Corbin, Jeremy Guthrie, Jason Marquis and Justin Masterson, none of whom has ever won more than 15 games in a season previously. Three of them entered the year with losing career records; Marquis was only two games over .500. Guthrie is 19 games under .500 counting his 2013 performance.

How many do you think will finish 2013 with at least 20 wins?

Answering that is nearly impossible, and as forward prognostication goes, entirely irrelevant. It is the mistake all too many fantasy owners make: In defending, say, Corbin's season-to-date performance, they'll argue, "But he's 7-0," when the far more compelling case to be made is that his FIP (Fielding Independent Pitching score) is 2.82, 15th among ERA qualifiers; his ground ball rate 50.3 percent, almost 5 percent higher than the league average; and his walk rate 7.5 percent, almost three-quarters of a percentage point beneath the league average.

Corbin's ability to avoid walks and induce grounders gives him a fighting chance once inevitable regression to the mean arrives, so the proper question about his future prospects is, what are his odds of maintaining an ERA beneath 3.50 and WHIP under 1.25? Any contribution in wins should flow from that answer; wins should not drive your projection, but rather be the byproduct of those ratios.

The same goes for Guthrie, whose continued reliance upon his sinker is largely behind his hot start; and Masterson, whose increased slider usage and effectiveness, especially against left-handed hitters, is behind his. Marquis would be in that group, too, if he didn't appear to have been entirely lucky (.231 BABIP, 80.3 LOB%); he has little to no chance at coming even close to those ratios going forward.

Here's why you shouldn't even bother to project wins: There isn't even a strong enough correlation between team influences -- run and bullpen support -- and wins as fantasy owners might believe. Examining statistics in the 13 completed seasons since 2000, only 55 percent of the time a pitcher reached the quality start minimum did he earn the win; 67 percent of the time he managed at least seven innings pitched with no greater than two earned runs allowed he won; and only 82 percent of the time he pitched eight-plus with zero or one earned runs allowed did he win. It's that 82 percent number -- approximately four-fifths -- that is shockingly low.

Breaking those numbers down further, let's illustrate what run support might have done for pitchers in each of these circumstances. Runs-per-game averages (the first column) are comparative to the league average; this year's Detroit Tigers, for example, would be in the "+0.50 or more" group thanks to averaging 1.01 runs per game more than the major league average (5.29, compared to 4.28).

Only at the extremes -- examples such as this year's Tigers, on the good side, or the Miami Marlins, on the bad -- did it appear that pitchers stood a significantly better or worse chance at notching the win, especially in that latter group, where there was effectively no difference on teams that averaged within a half-run of the league average. This "extremes" finding should come as no surprise; the shock should be that there's little-to-no difference between a merely above- or below-average team.

Now let's illustrate how bullpen support impacts the win column. In order to fairly judge relievers' performance, let's use ERA as our baseline, but also penalize relievers for the inherited runners that only starters left behind. As those aren't technically a reliever's responsibility, let's charge relievers one-half of an earned run for each of these "Bequeathed Runners" (the term Baseball-Reference.com uses) allowed to score, then recalculate that ERA. Again, we'll compare these "rERA" (Relief ERA) to the league's average.

Here's where it gets interesting: There is scarcely any relief impact upon starters' win potential except in the first two groups (quality starts and 7-and-2 games) in the instance of the game's truly worst bullpens. This means that, to use 2013 examples, only examples such as the Houston Astros (major league worst 4.80 relief ERA), Philadelphia Phillies (4.63 ERA and precious little setup help behind closer Jonathan Papelbon) and St. Louis Cardinals (4.54 ERA, nine blown saves) would be of significant concern.

What's more, look at those numbers from starts of eight innings or greater and zero or one earned run allowed: Teams with the worst bullpens actually protected the starter's lead (80.5 percent) more often than those with the best (79.3 percent). That's further evidence that good pitching -- not good run support or good relief work -- should fuel your wins projections.

For those lacking the time to do thorough such prognostication, here's a quick way to evaluate pitchers' year-to-date contributions, beyond the obviously valuable endeavor of examining their peripheral numbers (such as FIP, BABIP, LOB%, HR/FB% and batted-ball rates): Take our Player Rater numbers and extract wins from the equation. The charts below do this, the column on the left showing the top 30 starting pitchers (SP eligibles currently serving as relievers excluded) with wins included, the one on the right showing the top 30 excluding wins.

Here is where you can perhaps extract some underrated value, so long as you can live with a potentially diminished number of wins: Jeff Samardzija has the greatest differential (15 spots) between the two charts of any member of the top 50 starting pitchers; Jose Fernandez (13), Homer Bailey (9), James Shields (8) and Stephen Strasburg (8) are four others who haven't been as fortunate in the wins column as they could. Bailey and Strasburg in particular are puzzling inclusions, as both of them pitch for teams projected to rank among the game's best winners.

Using the reverse, the aforementioned Marquis, 45th among starting pitchers on our Player Rater, would rank only 76th (difference of 31) if wins were excluded. Similarly, Guthrie (32), Jorge De La Rosa (30), C.J. Wilson (15) and Lance Lynn (11) might perhaps be overrated due to their wins-inflated Player Rater standing. In the example of Guthrie, a poor strikeout rate contributes, while with De La Rosa, the nightly ERA/WHIP risk of calling Coors Field his home makes him risky.

A final thought: If unpredictability in the wins column still frustrates you, ESPN's League Manager affords you the ability to pick the categories you want, in the event you'd like to remove wins from the equation in 2014.

But that's a conversation for another time … oh, wait, it's one we've already had!


Note: Tristan H. Cockcroft's top 150 pitchers are ranked for their expected performance from this point forward, not for statistics that have already been accrued. For starter- or reliever-specific rankings, see the "Pos Rnk" column; these rankings can also be seen split up by position.