What miserable fortune Cole Hamels had last season.
On 17 occasions in 2013, Hamels pitched a quality start but failed to record a win; that matched an all-time, single-season record shared by seven different pitchers (most recently by Felix Hernandez, during his 2010 Cy Young season).
Conversely, six times last season, Tommy Milone's teammates "had his back," scoring him a win despite his inability to meet the quality-start standard (at least six innings pitched, no greater than three earned runs allowed).
Milone won 12 games and Hamels only eight, but no one out there could've said he or she would've preferred Milone on his/her fantasy team. In fact, using Player Rater data, Hamels would've finished 88 spots higher than Milone among starting pitchers had wins been stripped from the equation.
(As an aside and in the interest of historical perspective, 10 seasons ago, Kenny Rogers had a year of 10 "cheap wins" -- wins earned in non-quality starts. He was 18-9 despite a 4.76 ERA.)
It's examples like these that have fueled much criticism of the wins category, an amount that has increased in intensity during the past decade, and has resulted in many a league changing the category to compensate.
The category's detractors will quickly point out that run support and bullpen performance aren't individual skills, but both significantly impact wins, and in the specialization-rich modern game, those team influences carry increasing weight. To reward for wins, therefore, is to credit a player as much -- and often more -- for the name on his uniform than for his own ability.
Nevertheless, wins remain one of the 10 staple Rotisserie categories, and despite its criticism, it stands as part of the rankings formulation in countless fantasy baseball advice sources. If we're indeed in the midst of a statistical revolution, you might not think it, looking at most sources' rankings.
Now, to be clear: This is not a declaration that it's wrong to include wins as part of your fantasy baseball scoring system. Skill does influence the category, and there are arguments both for tradition -- wins were one of the eight original Rotisserie categories, chosen in 1980 -- as well as using categories that inject a bit of chance into the equation. It could be argued that our goal, as fantasy baseball owners, is to identify player skill, then use it to align our rosters with the best odds of success in more chance-oriented categories. Consider it one way to counter the annually increasing amount of information available; the more we can predict, the less "chance" there is in the game. Some people like the game with a larger amount of chance.
That decision is yours, and as you'll read below, as well as in my separate Rotisserie 6x6 categories story, there are compelling reasons to switch wins to quality starts. Absorb this data and make your own educated choice.
This space, however, serves as an answer to that age-old question: How does one ... er ... win "Wins"?
The evaporating starting-pitching wins market
Let's use facts to illustrate the specialization of the modern pitching staff: Last season, starting pitchers accumulated 1,658 of the 2,431 total wins recorded, or 68.2 percent. Twenty-five years earlier, in 1988, starters tallied 1,547 of the 2,100 wins (remember, there were only 26 teams back then), or 73.7 percent.
Wins are increasingly shifting into the hands of relief pitchers -- we'll get to them shortly -- both lowering the bar for starters in the category as well as increasing the amount of annual fluctuation within. To use another fact, and a historical benchmark, during a span of 12 non-strike seasons from 1980 to 1993 (1981 was strike-shortened), 56 pitchers scaled the 20-win plateau. Comparatively, during the past 13 full seasons from 2001 to 2013, only 40 managed to win at least 20. And only 16 of those happened in the past eight seasons.
One of the reasons for that is starters' inability -- or is it managers' preference -- to pitch deep into ballgames. Only twice in the 19 seasons since the 1994 player strike was the major league average of innings pitched per start greater than six (6.03 in 2011, and 6.06 in 1998). This puts many games' fates into the hands of relievers, and it increases the chances of a reliever ultimately getting the win.
It's as if the real game, despite the fantasy game's numerous critics of the category, is embracing the quality start's literal interpretation. Six-plus innings and three or fewer earned runs allowed seems a perfectly acceptable outing to modern-day managers, resulting in increasing randomness in individual wins.
How quality -- or dominant -- starts should drive your scouting
Quality-starts detractors claim that the measure results in a 4.50 ERA, which hardly passes the sniff test as far as they judge "quality." They need constant reminders that a 4.50 ERA -- that's exactly six innings of three earned runs allowed -- represents the minimum such qualification. And no outing that results in an ERA higher than that can possibly qualify.
Fans of the Fantasy Focus Baseball podcast might recall this fact, one of the first "Geeky Stats of the Day" of 2014: In 2013, there were 213 minimum-qualification quality starts (i.e., a 4.50 ERA); but there were 211 wins by starting pitchers who had an ERA of 4.50 or higher. It was really no less likely that your pitcher would've captured a win than a quality start with such an outing, and let's not ignore that 145 of those wins were by pitchers whose game ERAs were five-plus.
Let's break down all 2013 starts by results in wins and quality starts (and a combination of the two):
The win totals in those sub-two ERA games are expected, but it's both the low percentage of 1.00-1.99 ERA quality-start outings that resulted in wins, as well as the number of sub-two ERA games that failed that produce a win, that might take you aback. What's more, the number of 3.00-4.49 starts that resulted in wins seems higher than expected, especially if you delve further into those and realize that 67 of them were outings shorter than six innings in length.
All this points to quality of outing -- not specifically quality starts, but just the general quality of the performance -- heavily influencing wins (a no-duh conclusion, granted). It is for that reason that, instead of expending effort projecting wins, it might be smarter for you to project quality starts as you do your research, adjusting, naturally, for team and bullpen factors.
But, to give you an even greater head start, further research reveals a correlation between pitchers who go at least seven innings allowing two or fewer earned runs: In 2013, 744 of 1,205 pitchers who did, or 61.7 percent, won. Listed in the chart below are the season's leaders in these 7/2 outings:
Does length of start influence wins?
A pitcher's total innings pitched in the game in question does carry weight of its own in terms of his win potential. Again using 2013 data, consider:
• 68.9 percent of starts of 7 2/3+ innings resulted in a win.
• 48.5 percent of starts between 6 2/3-7 1/3 innings resulted in a win (with little variance between 6 2/3, seven or 7 1/3).
• 36.3 percent of starts between 6-6 1/3 innings resulted in a win.
• 24.4 percent of starts between 5-5 2/3 innings resulted in a win (with little variance between five, 5 1/3 or 5 2/3).
• 14.1 percent of starts shorter than six innings resulted in a win.
Why that 23rd out -- or 7 2/3 innings times three -- carries such significance is unclear; it'd make sense if there were a significant bump at an inning's conclusion, as there was a greater-than-10-percent bump when a pitcher finished six. Perhaps it's that recording it left four more outs in the hands of the bullpen, four outs often being a small enough number for a manager to trust handing that "extra" out to his closer (who is often the most talented reliever on the roster).
Here's a frustrating fact, though: Two of the four pitchers who averaged more than seven innings per start in 2013 failed to win as many as 15 games, as Cliff Lee (7.18 innings per start) won 14, Chris Sale (7.14) 11. And of the top seven in innings per start (minimum 20 starts), five fell short of 15 wins.
It seems that run support does trump workhorse ability -- Lee's Philadelphia Phillies finished 26th in runs scored in 2013, Sale's Chicago White Sox 29th -- which is why you shouldn't put too much additional weight on high-innings guys.
Last season, 773 of the 2,431 total regular-season wins recorded (or 31.8 percent) were recorded by relief pitchers. That might not sound like much, but those 773 set a new single-season record, and that 31.8 percent also represented the largest such percentage in any season since World War I.
Fantasy's proponents of wins as a category might argue that this represents the great starter-reliever equalizer. This ignores, however, that the written rule regarding relief wins is subjective in one regard (the portion that grants the official scorer the right to assign the win at his/her discretion if the starter fails to complete five innings) and not in another (the part that hands the win to the pitcher who was in the game at the time his team captured its final lead). Pitchers who benefit from the former are notoriously difficult to predict; pitchers who fall into the latter don't even have to pitch effectively in order to win.
To the latter point, consider that 62 of those 773 relief wins in 2013 came by pitchers who suffered a blown save, then subsequently had their teams rally to recapture them a lead. Fifty-eight of those 773, meanwhile, came from relievers who posted a 9.00 ERA or higher in the given outing, and 85 were recorded by relievers who allowed at least one earned run.
Just as a matter of perspective, on 95 occasions last season, a relief pitcher pitched at least two hitless, scoreless innings with three or more strikeouts without earning a win for his efforts. Was it necessarily fair that those 85 who afforded runs won, while these 95 did not?
Still, these relief wins count, so let's examine the circumstances most likely to fuel them. First, let's break them down by point of the game the reliever entered:
No one should be surprised by the higher percentages in the first five innings; these play to the "starter-went-less-than-five" rule, which often results in the official scorer handing the win to the pitcher who recorded the final out of the fifth inning (hence the 10.0 percentage in that frame). The 21.0 percent rate in extra innings is also no surprise, considering its sudden-death nature.
But it's that small differential between the seventh, eighth and ninth innings that might catch your eye, and it points to closers enjoying a small hint of a wins advantage. That could be the product of managers' occasional -- and wise -- decision to summon the closer into the ninth inning of a tie game at home; and sure enough, 108 of those 153 ninth-inning wins came in home games.
Now, let's break relief wins down by innings pitched in the outing:
Again, it should come as no shock that the round numbers -- exactly one, two, three, etc., innings pitched -- result in a greater number of wins, because of the likelihood that those were complete-inning outings (i.e. the reliever both began and concluded the frame); the reliever's team then had an opportunity to bat and claim him the win. Taking that into account, there is practically nothing here to indicate that length of outing grants any sort of relief-win advantage. When 55 pitchers can record a single out yet capture the win, an identical number as the 55 who earned the win with at least 2 1/3 innings pitched, it is clear that there is no correlation between distance and win probability.
In short, expect a minuscule wins advantage from your closers, but beyond that, understand that relief wins are inherently random.
Run support and matchups
It's the final topic of the bunch, not because it lacks importance, but rather because of how obvious it is that a team's level of run support, as well as the relative strength of a pitcher's opponent, fuels the wins category.
Last season, three of the four pitchers who won 18 or more games pitched for 93-plus-win teams, and the fourth, Jordan Zimmermann, pitched for the 86-win Washington Nationals. All but four of the 16 pitchers to win 15 or more games, meanwhile, pitched for teams that finished in the upper half of the majors in terms of runs scored, seven of them pitching for top-five squads.
Looking at it from a team's perspective, the chart below ranks each team, on the left by percentage of its starting-pitching wins that were also quality starts, and on the right by the total number of wins earned in starts that failed to meet the quality-start minimum.
There are outliers on these charts -- the Los Angeles Angels stand out on both sides -- but it's clear that being a starter on a light-hitting team means a smaller margin for error (again, a no-duh conclusion). But what's interesting about these findings pertains to fantasy owners in daily leagues: The portion to the right shows that, when considering pitchers of lesser quality, it's not a bad idea to "chase wins" by plucking ones who pitch for the most potent offensive teams. Felix Doubront is an excellent such example, a pitcher who registered ERAs higher than four in each of the past two seasons, but also one who won 11 games in each of those years. Knowing that Doubront was getting a starting assignment on a given day, it might've been smart to pluck him off your free-agent list and stream him, just because he had better odds of earning you a win thanks to the offense backing him.
Again using matchups -- this is something especially important to those of you who utilize said "streaming," or maximize-your-individual-matchups, strategy -- the chart below shows the number of wins recorded by each team's opposing starting pitcher, as well as that team's rank in terms of runs scored in 2013:
This pairs with the Doubront example, as the chart shows that picking the right matchup is every bit as important as simply chasing pitchers on the most successful teams. What's important is that it's not just the matchup -- it's also the pitcher's team -- as the two charts above show.
Does this mean you need to replace wins with quality starts in your league, to avoid both the randomness inherent to the category as well as the likelihood that more fantasy owners will lean on streaming to fill it? Or might it mean that you merely need to adapt your strategy as wins become increasingly difficult to project?
It's your choice. But one thing's clear: Recognize the game's a-changin', and both your strategy -- and your taste in categories -- need more careful consideration.