Pitcher wins are both the most exciting and most frustrating category in fantasy. One day, they vindicate a seven-inning, two-run effort. The next they bail you out of a five-inning, six-run stinker. The one after that, a lack of run support ruins an exemplary eight-inning, 10-strikeout shutout.

They’re erratic, they’re wildly unpredictable, and they’re an unavoidable part of fantasy life.

They’re also a fantasy strategist’s greatest opportunity.

In an age where we espouse the virtue of milking every bit of value from every corner and under every unturned stone, wins remains one of the least explored categories—one still open to discovery and exploitation.

And where there’s opportunity, you can bet that smart, championship-caliber owners will take advantage. And there’s an opportunity here, readers. Where lesser owners shy away from the category, ignoring the wins column as if it were called “runs,” you can take advantage.

First, the hard part: How do you weave through the inconsistency and figure out how many wins my pitcher will rack up this year? As a statistic that is fraught with unpredictability, this must be easier said than done, right?

Wrong!

Luckily for us, FantasyPlayerRater.com has published a wins formula (follow this link, first article down) that can help you with that first step. It calculates starting pitcher wins per game started, taking into account a pitcher’s ERA, run support, bullpen strength and a couple of other variables.

ERA, obviously, is the most important factor, followed by run support. Bullpen strength, AL v. NL, and innings pitched per games started are also important, but those first two dominate the landscape.

(A side note: I hope by now that THT, Baseball Prospectus, and Fangraphs have fully debunked any of those “this guy gets more run support than this guy in the same rotation because…” theories. To those who still believe: Run support has nothing to do with who’s on the mound. You can estimate run support by dividing the number of runs scored in a season (or expected runs scored) by 162 games.)

I would like to caution anyone against understating the value of run support—or team quality in general. I’ve come across many who argue that wins are so volatile that they should be ignored altogether. I would caution against that sentiment. So, just how important is team strength when you’re determining expected wins?

In a one word: very. In two words: very, very. Get the picture? It’s extraordinarily important. In fact, a huge amount of a pitcher’s value hinges on this statistic.

To illustrate the point, let’s take Yankees ace C.C. Sabathia 2011 season and see how many games he would have won had he been on each team in the majors:

{exp:list_maker}Sabathia’s profile in 2011: 3.00 ERA, 7.192 IP/GS, American League pitcher.

Yankees’ average run support per game: 5.352 (867 runs / 162 games)

Yankees’ bullpen ERA: 3.12

{/exp:list_maker}

The results, with Sabathia holding constant at a 3.00 ERA:

By the formula, Sabathia was expected to lock down just shy of 20 wins. This is encouraging: He won 19, so we’re within one win of the actual result.

With the results in hand, I would like to draw the reader’s attention to three main pieces.

First, column four illustrates the expected number of wins for Sabathia on each of the 30 teams.

Building on that, column two reports how many points above or below average that win total is according to FantasyPlayerRater.com’s roto-points calculator. Said another way, this is the expected affect C.C. Sabathia will have on your final standings in the category. So, if you were at six or seven points at the end of the year in wins (the middle of the pack in a 12-team league), adding Sabathia would have brought you up to 8.5-9.5 points in wins. As fantasy owners can tell, that 2.5 points is a huge margin and can spell victory, defeat, or standings position in any given season.

The third column depicts how many points a team would gain/lose in the standings if Sabathia switched teams from the Yankees to a new home.

Looking down the list, we can see that the best team for wins was, unsurprisingly, the Yankees, followed closely by the Red Sox. The worst team for wins was the Astros, followed closely by the Mariners. On either of these two teams, Sabathia would have lost more than six wins—a staggering sum that would have accounted for a loss of more than 2.5 points in 12-team leagues, according to FantasyPlayerRater.com’s ratings.

That’s a huge number. Is anyone out there worried that they may have fallen short of a championship because they chose Seattle’s Felix Hernandez and the pathetic Mariners offense over the vaunted Yankees? I know I would be. In my main league last season, this exact scenario played out: The first-place team finished with 90 points and owned Sabathia, while the second place team finished with 88 points while owning Felix Hernandez. Could a simple trade have brought victory to the second place team? It’s a frightening thought.

But now we know! And now we can take advantage of this easily exploitable category and start winning championships. Now that the information is out there, go hit the free agent and trade market and bring home some heists! With half a season left, what are you waiting for?

Detroit Michael said...

Of course, Sabathia version 2011 would not have had a constant ERA regardless of the ballpark factor and defense behind him. Still, it’s a nice illustration.

Mike Silver said...

Agreed.

The improvement in ERA and ballpark factor would also factor into the win column—as well as the overall value calculation, but it in terms of a quick analysis, its there to illustrate the main point.

KY said...

I have been operating on the win prediction method of pitcher efficiency also. That would make another good article for predicting wins. The less pitches it takes you to get outs the deeper into games you get. In a deep NL league guys like Aaron Cook or Nick Blackburn have served well back when they were throwing better. They would get more wins then K pitchers on their same team by lasting longer on the average.

Mike Silver said...

@KY

I really like that idea. I suppose you create a regression to calculate IP/GS that takes into account pitcher efficiency. That would certainly take out some of the uncertainty in predicting IP/GS, whereas up until now I have always just taken the shortcut and used career average.

Abdoozy said...

I’m hoping this is satirical analysis.

Sabathia started 33 of NY’s 162 games, roughly equal to a 1/5 split of the team starts, as you would expect of a healthy ace. NY won 97 games. 1/5 of 97 = a projected 19.4 wins. CC won 19.

Seattle won 67 games. 1/5 of 67 = 13.4 games. And, what do you know, Felix won 14 last year.

Wow, starters on teams that win more games get more Ws. How about that! I’m a Sabermetrician!

KY said...

Well, I prefer pitches as it gives credit to those who pitch deep into games efficiently, not just pitch deep into games. Also, this is a bad stat for individual seasons as it contains hits and defense. Usually the guys who can do this well are ground ballers and fly ballers which adds more value then you previously might have suspected a ground baller with low K’s to have.

Fun example, in the second half last year, JA Happ was the least efficient starter with only 14.7 outs per 100 pitches in the second half. A full out worse then any other starter. This year he’s up to 16.7 which is still bad. The best pitchers are usually about 20.3 or so. RA Dickey was the 6th most efficient pitcher in the second half last year of our NL only league. Its fun to note that, if you figure 100 pitches per start, basically great pitchers can get an extra 4 hitters out per game then bad pitchers.

Another article might be the effect of team defense on wins. A great defense gets a pitcher an extra .5-1 out per game too, which effects wins. Maybe a 5th column on your chart for team defense? And a 6th for your IP/start idea? The projection engine guys may already do this too as they all predict wins.

Mike Silver said...

@KY

Those are some very good points and should definitely be factored into any regression on the topic.

Going further, I would say that total pitches per game started would also be a good stat to track to account for things like pitcher stamina or manager usage patterns. For instance, Dusty Baker likes to burn out his starting pitchers, so those pitchers now throw 10 extra pitches and thus, 2 more outs.

I must confess I’ve never studied the topic, but I imagine that certain pitchers throw more pitches per start (whether due to stamina or manager preference), which would impact the outcome.

Mike Silver said...

@Abdoozy

If you want to say that better teams win more games, I don’t think you’d find a single person in the world who would disagree with you.

If you’re trying to say that you can take a team’s wins and divide by 5 to get an SP’s projected wins, you’re sorely mistaken. A.J. Burnett’s 5.16 ERA gives him 13.4 expected wins. By your formula, the Yankees would have won 67 games that season—30 short of their actual total.

The point of the piece is that of being able to take control of the wins category and get a real estimation of SP wins before they happen—not taking a team’s win total after the season has ended and dividing it by 5.

Check the formula:

Pitcher Wins per game started = -0.0594893 * + 0.0889605* <Average Runs Scored per Game by Pitcher’s Team> + -0.0279382 * <Team’s Bullpen ERA> + 0.0528664 * <Innings Pitcher per Game Started> + <0.0132315 if American League pitcher, 0.000 if National League Pitcher> + -0.0020577

Far more accurate than dividing team wins by 5.

Abdoozy said...

AJ Burnett wasnt close to being an ace last season. That rough estimate doesn’t work for down-roster starters, for obvious reasons.

Of course you can’t assume that ANY starter gets 1/5 of any team’s wins. But it’s pretty easy to assume that CC, as the Yankee one, is going to probably have more wins than Felix. Your research confirms this.

I’ve used a strategy of trying to get good starters on teams projected to win a lot of games every year. It works out pretty well, because those starters often have good numbers across the board, and the Ws are a nice bonus.

Mike Silver said...

I think there’s a miscommunication on the main point of the article. Obviously, Yankees pitchers are better positioned for wins than Mariners starters. That’s indisputable.

The point I’m aiming for is that wins need to be given as much attention as the other categories and the way this is done is by using projection formulas to get a better sense of who’s better than who and by exatly how much.

Otherwise, taking shortcuts and over-simplifying the wins category means you miss out on those 4 or 5 wins among the 5-6 SPs on your staff that costs you those 1-2 crucial points in the standings.

Ruff said...

Your entry inspired me to trade Teixeira for an injured CC. This is dynasty league, and I HATE trading for pitching help but W’s have evaded my staff and it was time to say goodbye to my longest tenured player (owned Tex since his arrival). Thanks for the informative read.