Seeking Alternatives to Wins

Enjoys long walks at more appropriate places in his fantasy baseball league’s standings. Must love dogs. (I do like me some Diane Lane, one of the more underrated Hollywood beauties of her time, in my opinion.)

This is kind of like the personals section of your local Craigslist, minus the creep or latent possibility of disease, death or need for a restraining order. There’s also practically a 0% chance that this entry ends up as a prompt for a category on the hilarious Comedy Central program “@midnight.”

I’m tired of wins. Not just because the stat has quote-unquote screwed me out of a rotisserie league title, although I’m sure it has at one time or another. Not just because of all the good arguments for why it’s a terrible statistic for individual pitchers. Not just because I’m tired of hearing people complain about why it’s a terrible statistic for individual pitchers. I’m over that stuff, for the most part. I’m just tired. And kind of bored.

This is an age-old problem. Scott Spratt most recently discussed the importance of the dissociation of wins from pitcher production. He referenced Jeff Zimmerman’s preseason work on the subject. We don’t like to use projections for the statistic to appraise future pitcher performance.

Some commissioners have attempted to address the conundrum with the substitution of quality starts (QS) or wins plus quality starts (W+QS). I’ve played in leagues with one or the other. I don’t recall a complaint about either of them. One of those two is clearly better than wins alone, as Community Research contributor Sam Horwich-Scholefield concluded. Neither is perfect, of course. I don’t recall disagreements when someone has inevitably asserted that three earned runs conceded in six innings is of great quality. It’s a binary statistic, just like all the other non-rate stats.

Some head-to-head leagues subtract points for or include the category of losses. Talk about rough for the 2014 Jeff Samardzija owner. Keep running him out there, brave soldier. Don’t let the fear win.

Perhaps I’m asking for too much when I seek an alternative to quality starts, then. Imperfections can accentuate beauty. But I’m asking anyway. Can we do better?

This substitute should:

  • more accurately gauge a pitcher’s contribution to his team’s chance to win a particular game
  • not be a rate
  • be exclusive to starting pitchers; saves, holds and potential alternatives to those (good examples of which I’ve seen here a time or two) are exclusive to relievers
  • avoid duplicity, or at least excessive duplicity (which then makes the quality of duplicity subjective, but whatever)

Anything else?

On the podcast, I suggested two possibilities, each of which would be a modification of an invention for said purpose: Pure Quality Starts (PQS), manufactured by Baseball HQ; and Bill James’ more well-known Game Score (GmSc), for which Baseball HQ created their simplistic alternative. Each is intended to be more of non-fantasy measure, but doesn’t either fit the bill? Has anyone at FanGraphs developed an alternative to them? If so, I’m sorry that I missed it. While I’m asking, can I be fired for either of these suggestions?

Again, this would be a cumulative category. Some might not be in favor of a category that results in a positive contribution for an awful outing on the mound, such as a PQS 1 (disaster) or a GmSc like Edinson Volquez’s 12 on Wednesday at home against the Cincinnati Reds. It’s just semantics, though. The scaling of one of those algorithmic stats could improve the aesthetic appeal of the category to the community.

Neither of these is perfect, either, which is why I suggest modifications. That and doing so may help me to avoid copyright infringement. Anyway, each of them is overly redundant for fantasy purposes. To implement them could translate to substitutions for ERA, WHIP, strikeouts or all of them.

At this point, I may have made it too complex. The consensus best alternative might already be out there. QS or W+QS might be the best solution because it’s simple and not superfluous. I’d just like to see something like it included in mainstream leagues. What do you think, Tout Wars? Is it time to start a movement?

I think the qualitative nature of a modified PQS or GmSc is really appealing and would add an interesting dimension to rotisserie races, too. It’s fun to contemplate.




Print This Post

Nicholas Minnix oversaw baseball content for six years at KFFL, where he held the loose title of Managing Editor for seven and a half before he joined FanGraphs. He played in both Tout Wars and LABR from 2010 through 2014. Follow him on Twitter @NicholasMinnix.


63 Responses to “Seeking Alternatives to Wins”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Av says:

    What about Innings?

    Vote -1 Vote +1

    • MDL says:

      That’s not a bad start, but you allow inning-eaters that just aren’t that good.

      I was thinking something like Outs – Bases Allowed (or something like INN*3 – (1B + 2B + 3B + HR + BB)) to be more generous to pitchers who also prevent baserunners. Total Batters Faced – Bases Allowed might work as well; the greater the number means more people attempted and failed to reach base.

      Vote -1 Vote +1

      • Jacob says:

        This is starting to just sound like a points league.

        Vote -1 Vote +1

      • MDL says:

        That should read “1B + 2*2B + 3*3B + 4*HR”. Anyway it’s hard to find a site that tracks doubles/triples allowed, but here’s a rough 2014 leaderboard for Outs – Total Bases Allowed:

        Rnk,PLAYER,O-TBA,WAR Rnk
        1,Johnny Cueto,221,4
        2,Felix Hernandez,221,8
        3,Adam Wainwright,195,6
        4,Julio Teheran,193,7
        5,Sonny Gray,180,23
        6,Garrett Richards,177,19
        7,Masahiro Tanaka,172,1
        8,Dallas Keuchel,172,2
        9,Kyle Lohse,172,30
        10,Scott Kazmir,170,10
        11,Jeff Samardzija,165,14
        12,Michael Wacha,164,26
        13,Mark Buehrle,162,3
        14,Yu Darvish,162,5
        15,Charlie Morton,161,39
        16,John Lackey,160,15
        17,Corey Kluber,160,31
        18,Jason Hammel,156,9
        19,Tyson Ross,154,40
        20,Jonathon Niese,152,24

        Not bad.

        Vote -1 Vote +1

  2. kevinthecomic says:

    I am in a Yahoo! league that uses QS as a stat. It is definitely superior to wins, but has a pretty significant flaw in it. I will demonstrate with an example. Let’s say a pitcher gets through six innings and allows three runs or less (a QS) but then the manager brings him back out for the 7th and the pitcher proceeds to puke all over his shoes (and end up allowing more than 3 runs). Now, as I understand what QS is trying to measure, this result is on the manager, not on the pitcher. The pitcher delivered a quality start but the manager blew it. In fact, BP has a manager stat called blown quality start (BQS). In our league, in this example, my pitcher would not get credit for a QS, even though he did in fact deliver a QS that was subsequently blown by the manager. I am having a hard time getting my league-mates to understand this, let alone trying to get someone at Yahoo! to figure it out.

    Vote -1 Vote +1

    • my jays are red says:

      It is also the pitcher/defense’s fault for giving up the runs. There are obviously issues with QS, but this seems like pinpointing a fairly minor problem that obviously should not be identified as a QS.

      Vote -1 Vote +1

      • ImDownWithOBP says:

        I agree with my jays. Maybe the manager made a mistake sending the pitcher out there again, but in the end t was the pitcher who gave up the runs. If a pitcher goes 6.1 and gives up 6 runs (3 in the final .1 inning) that most definitely is not a QS.

        Vote -1 Vote +1

    • sugarbush says:

      I couldn’t agree more. I hereby propose QSTSI (quality starts through six innings). If a starter comes out and gives six innings of a QS, that gives their team a decent chance to win. The SP shouldn’t be penalized for having a manager run them out there until they get too tired. I owned David Price for a few years in a league, and this seemed to happen all the time with him. For example, TB would be up like 8-1 after six, Maddon brings Price out to pitch the seventh and he proceeds to shart all over himself (say he gives up 3 more ER). He gets the W, TB wins the game because Price pitched well enough to keep them in the game, but no QS for fantasy. Awful.

      Vote -1 Vote +1

      • kevinthecomic says:

        Funny that you bring up David Price. I have him this year and Maddon is doing the exact same thing!!!

        Vote -1 Vote +1

      • zurzles says:

        i seriously can’t tell if either of you are being serious or not. you know ERA and WHIP are also affected by bad managerial decisions, right?

        Vote -1 Vote +1

      • zurzles says:

        after years in a QS league, i hate QS and prefer W. don’t really care if QS means more (in leagues where RBI and R count, that’s not really saying anything), there’s still nothing more frustrating than your pitcher going 5.2 scoreless or 9IP/4ER. i prefer the randomness of wins, honestly.

        ideally, i’d like to use game score or PQS but make them binary — instead of adding up totals, set a target score and give one “point” if they reach that target. a game score of 60 (or whatever’s good enough to be called “good”) or a PQS of 3 or 4. kershaw’s had 8 games with a GS of 60+, so he’d be worth 8 of whatever you’d call that stat. we keep QS and Ws around cause we like that binary “pitcher done good vs pitcher didn’t done good” stat. i do anyway.

        Vote -1 Vote +1

      • Mark L says:

        The Yahoo league(s) I’m in just pit the two sets of pitchers against each other – whoever has the highest Game Score for the week gets 1 point.

        This is perhaps a little too basic? I don’t know. I’d still be losing, no matter what score we used :(

        Vote -1 Vote +1

    • Ben says:

      Wouldn’t a pitcher also lose a QS if they went a full 9 innings but allowed 4 runs? The 3 runs/6 innings requirement seems pretty arbitrary in this scenario. I’m a big fan of QS over W, buts it’s definitely still a flawed stat.

      Vote -1 Vote +1

    • JMo37 says:

      That’s the funniest thing I have read, Kevin the Comic. You are suggesting that your pitcher got hit because the manager brought him back out.?.
      These stats exist and are accepted for a reason. Sure, there are always examples of ‘why it could be better’, but a category blaming the manager…..that’s simply not going to fly.

      Vote -1 Vote +1

      • kevinthecomic says:

        Like I said in my original post, it isn’t just me. Baseball Prospectus uses the stat of Blown Quality Start (BQS) to measure the performance of the manager.

        Vote -1 Vote +1

  3. Ryan says:

    How about Innings Pitched? The problem with QS, gamescore, etc is that they correlate too highly with the rate stats we already count. It seems to me the real thing you try to capture with a counting stat for SPs is that you’re taking on more risk in ERA/WHIP by having a SP instead of an RP.

    Vote -1 Vote +1

    • Jason B says:

      I’m not a huge fan of IP…it’s not a bad proxy for quality at times, in that the best pitchers are often reflected near the top of the IP leaderboard, but a high inning count may also showcase a lack of organizational depth as much as a pitcher’s quality. (As an example, a really awful team trotting out a very bad SP4 or SP5 for 30+ starts only due to lack of alternatives in-house or in the minors, or an unwillingness to spend to acquire a suitable replacement.)

      a/k/a the Jeff Suppan corollary. (Randy Wolf? Jose Lima?)

      Vote -1 Vote +1

    • Brad Johnson says:

      IP can’t work for any normal roto league. Only a non-serious owner misses the innings cap. It can “work” for H2H, but that’s just a measure of volume.

      Vote -1 Vote +1

  4. Kevin says:

    I think the best so solution is a tweak to the QS stat itself to extend it to 7 IP. I don’t know how one would go about altering the official definition of a stat though.

    Vote -1 Vote +1

    • buzzo says:

      Somebody would still complain that it isn’t “quality” enough. From my experience, people who don’t like QS get waaay too hung up on trying define something subjective like “quality”. If the name was “start that met a reasonable minimum threshold of both runs allowed and innings pitched” there would likely be less whining over a start that met the bare minimum. QS is a great stat to use for fantasy purposes.

      Vote -1 Vote +1

  5. David says:

    Wins are OK. They are quite random, but that is one of the things that makes fantasy feel like real baseball. Wins care not for your sense of justice and fair play.

    We all know that some guys will get fewer than they deserve (like Felix, and to a lesser extent Kuroda and Lackey). We know some middle reliever will have 10-12 when all is said and done. Make a sacrifice to the fantasy gods and hope for the best. Accept that some things are beyond our control and embrace the anarchy that Wins introduce into our lives.

    +5 Vote -1 Vote +1

    • Cuck City says:

      Koji Uehara blows a save and got a win yesterday

      WELL DESERVED WIN

      Vote -1 Vote +1

      • David says:

        As I said, Wins are almost completely random. Lackey owners are frustrated, Uehara owners got an unexpected (undeserved?) boost.

        (BTW, he entered a tied game so he didn’t blow a save. He was in line for the loss until the bats bailed him out though)

        Vote -1 Vote +1

    • ImDownWithOBP says:

      The point of fantasy baseball (aside from having fun) is to prove that you are better than friends, so when I wave my trophy in their faces, I want them to know that I was objectively superior to them. I don’t want the randomness of wins, and my competitors subsequent jabberings about luck, to taint my personal championship parade. I want a statistic with as few variables outside of my control as possible. Variables open the door to arguments about luck, and arguments get loud, and I cannot afford the yelling, because my mom has to get up for work in the morning.

      +7 Vote -1 Vote +1

      • Lowest NERD Scores NA says:

        I know how to solve this problem. Beat your friends year in and year out. That way there is no arguing, or debate even. You are simply the God of Fantasy amongst your friends

        Vote -1 Vote +1

      • David says:

        Most of the standard categories have a measure of luck involved. RBI? Runs? HR (Coco Crisp had 20 last year, through no amount of skill of his fantasy owners)? WHIP? SAVES?

        You can try all you want to control the environment, but that is largely an illusion. Strasburg just gave up seven runs to Milwaukee, with MIL sitting most of their lineup.

        Vote -1 Vote +1

  6. Cuck City says:

    Average Game Score – making a rate of out it, would be best IMO

    Vote -1 Vote +1

    • Cuck City says:

      With a minimum IP / Games Started of course

      Vote -1 Vote +1

    • Roger says:

      How about researching what’s an average game score and subtracting that from the game score? So if in an average outing an SP registers a 50 game score, Edinson Volquez’s 12 would register as a -38.

      Vote -1 Vote +1

      • Cuck City says:

        so Gmsc+ basically

        Vote -1 Vote +1

      • novaether says:

        It’d be better to find out the replacement level game score and subtract that. This way streaming should, in theory, add 0 points. If you subtract out average game score then the best strategy would be to pitch Kershaw once and throw in relievers to pick up 4/5 categories.

        Vote -1 Vote +1

  7. Walks with four balls says:

    Speaking of ignoring wins. I’m having trouble managing the back end of my rotation and wanted to see if it would be more beneficial to pick up a guy like Wade Davis or another set up man. I’m tired of the carousel ride. I can also pick up another closer like McGee, but a dominant set up guy might be better for my ratios. What do u think?

    Vote -1 Vote +1

  8. Capt. Hook says:

    How about: A pitcher goes at least five innings. If his team is tied at the end of the inning in which he departs, he gets 1/2 point. If his team is ahead, he gets 1 point.

    It would capture the “head-to-head” spirit of wins, but not be ruined by things that happen after (which the pitcher has no control over). If you have two pitchers going against each other, you could theoretically get 1 1/2 or 2 points.

    It would not fix the Jeff Samardzija problem of no run support, but those owners are already rewarded by positive contributions in ERA and Whip (and often K).

    Vote -1 Vote +1

  9. Mix says:

    I’m a big fan of QS.

    We tried IP one season in a league with daily changes. Ended up seeing some awful pitchers being tossed out there as just pure innings fodder and random strikeout accumulation. That’s the opposite effect of your stated goal of rewarding a quality line-up.

    Vote -1 Vote +1

    • ImDownWithOBP says:

      This is exactly what has happened the past couple years in my 2 H2H leagues. If someone had a SP get blown up, he would surmise that he could not win ERA/WHIP that week and would then start throwing crappy starters from the wire out there in an attempt to take IP, K, and QS. I always felt IP creates a perverse incentive to seek out quantity over quality.

      Vote -1 Vote +1

  10. nj says:

    The problem with quality starts is that it puts 6 innings 3 runs on the same footing as a complete game shutout. Wins actually reward the better start in theory because in a vacuum the better the start, the better the chance for a win. And lets not fool ourselves into thinking that wins are the worst stat there is because it relies to heavily on outside forces. It just creates a strategy when you draft that requires you to take the teammates into consideration, that is involved in many stats (RBIs, RUNs etc.) Instead you can weight the different types of starts. 1 point for QS (6 innings 3 or less runs) 1.5 points for good start (7 innings 2 or less runs) 2 points for excellent start (8 innings 1 run or less) 3 points for a complete game shutout.

    Vote -1 Vote +1

  11. Running Frog says:

    One question that should be asked is do you even need a replacement for it? Go back to your typical end of year standings and look how often the result of the Ws stat actually affects your leaderboard. It’s pretty much never. The teams that do well in wins are the same ones who do well in Ks and the ratios.

    I’ve been playing fantasy since ’82, and I was really glad when leagues started tracking Ks as it added a lot more strategy to playing the game. But you could have just dropped Ws when you did it with few ill effects.

    Vote -1 Vote +1

  12. Ned says:

    I realize it’s a rate stat, but the best one is IP/Start. It helps bridge the gap between super valuable, real life pitchers who get ahead and get weak outs vs the guys who throw 6-7 innings and get super high Ks but can tax a bullpen. Think of it as the Maddux stat.

    Vote -1 Vote +1

  13. Remus says:

    I don’t see the problem with it, you try to get the best fantasy pitcher, not the best real life pitcher, because those many not be the same person.

    If you have a problem with it, why not have a problem with the catcher eligibility on people who obviously wont catch anymore, or the double SP/RP eligibility?

    I’d rather play with QS, but wins make sense from a pure fantasy perspective when bundled with all the other stuff in which if differs from real life baseball.

    Vote -1 Vote +1

  14. FeslenR says:

    why not just use both? I play in a league where both QS and W are stats, and it works just fine for me. After all, if a pitcher pitches really well but his bp or offense doesn’t back him up (almost all Met starters(, then both qualify as a way to categorize the starting pitcher of the day.

    Vote -1 Vote +1

  15. Metsox says:

    We use QS+CG, you get an extra point for a going the full 9 innings (or occasionally the full 8 on the road).

    Works well for us….

    Vote -1 Vote +1

  16. Ian says:

    Game Score seems like a poor idea, because most of the factors (high K’s, few baserunners, few runs) are incorporated into existing stat categories.

    I don’t mind the win. It is random, and thus adds some fun. And it may go against everything else in the stat line for the pitcher for the day.

    Vote -1 Vote +1

  17. David says:

    We use (W+QS-L) in our league, and it works pretty well. Though the 8IP 4ER non-QS always sticks in my craw…

    Vote -1 Vote +1

    • Jason B says:

      I dunno how “quality” that 8 IP, 4 ER start is anyway. It’s a 4.50 ERA, which is well north of league average nowadays.

      Vote -1 Vote +1

      • nj says:

        6 innings 3 runs is also 4.50 era

        Vote -1 Vote +1

      • atoms says:

        Exactly. If 6 innings of 4.50 ERA pitching is “quality”, how is 8 innings of 4.50 ERA pitching somehow NOT quality?

        Vote -1 Vote +1

      • lowcountryjoe says:

        It sure the hell would put a team in a good position to win the game, though. The stat ain’t perfect but it is better than the W. Also keep in mind that this would be the outer boundary of eligibility. Further keep in mind that if a real team knew they could get that exact result from a starting pitcher every time out, that pitcher would make $18+ million a season and would matchup with the opposing team’s #2 SP every single time and would rarely lose a game.

        Vote -1 Vote +1

  18. shibboleth says:

    Big fan of PQS, and that would be my #1 choice. Until Yahoo roles it out, though, I’d settle for QS+W. Although, Baseball Forecaster did a study about the difference between those two, and IIRC, it wasn’t as big a difference as you’d expect.

    Vote -1 Vote +1

  19. Bobby Bob says:

    While we’re on pitching categories, I’m in a league that ratios Ks with Ks/9 as a category as opposed to just total strikeouts. I think this sways the balance in pitching to almost favor relief pitchers as ERA, WHIP, Saves, and Ks/9 are all usually going to favor relief guys. What do you guys do in your leagues for strikeouts?

    Vote -1 Vote +1

    • Rob says:

      We use K/9 but balance it with QS and IP so that relievers aren’t favored too much. Otherwise, in a league that uses QS or Wins, ERA, WHIP, Saves and K/9, there would be no incentive to roster any starters at all except for the very elite ones. An all-RP roster would be very good in that league. Conversely, adding IP favors quantity of starters of quality and favors streaming a bunch of bad starters. Adding K/9 with IP balances this nicely.

      Vote -1 Vote +1

    • Brad Johnson says:

      If your league has an innings cap, you’re already using K/9. It’s just non-obvious and you’re forced to hit the innings cap.

      Vote -1 Vote +1

  20. Jake says:

    Bottom line the criteria for a win has to change and it should be this, If a starting pitcher completes five innings and leaves with the lead no matter how many times that lead changes hands should his team go on to win the win reverts back to that starter period. He did his job and outperformed the opposing pitcher and should be awarded with the win. How many times do we have to see someone from the bullpen blow a lead and then be awarded the win just retarded.

    Vote -1 Vote +1

    • aklang says:

      I agree with Jake, or it could even be something as simple as “team wins start” with a 5ip minimum. Then you could also argue that as long as the pitcher has a lead or even a tie that they have done enough to earn the win.

      Vote -1 Vote +1

  21. jdbolick says:

    There is a pervasive misconception about fantasy baseball that it either is supposed to or else should be a simulation of real baseball. The reasoning goes that because X isn’t that important for real baseball, that categories should be changed in fantasy baseball to diminish the value of X. But fantasy baseball isn’t a simulation; it is a game.

    The next response might follow that games are supposed to be about skill and there are better categories to choose than wins in order to reflect skill in identifying pitching quality. While that is true, I would argue that it is essential for games to also be “fun.” It can be frustrating to see a bullpen blow a lead for your starter, especially in September when you’re one win away from a critical half point. But who hasn’t gotten a smile on their face due to an unexpected win from a middle reliever? There is an element of randomness in the wins category that makes it entertaining, especially since it still does correlate with pitching quality, just not as strongly as some other possibilities. That frustration of losing something undeservedly and the joy of finding free money makes fantasy baseball better to me.

    Or if you don’t find my whimsical approach to the game compelling, consider that ERA and WHIP already provide some measure of pitching quality as well. So even if you unjustly lost a win because of some other player’s failing, you still get substantial benefit from a good starting performance. By switching from wins to quality starts, are you not making the entire category redundant given that it is based on ERA? So what would possibly be the point of that?

    I’ve been writing about fantasy football and fantasy baseball for over a decade, and any fantasy football aficionado should be able to relate the chaos and confusion of all the various scoring systems in different leagues. At a certain point you’ve stopped making the competition about picking players and started making it about understanding the rules better than your opponents do. Maybe others disagree, but to me that’s not fun. I want to win leagues when my competitors and I all had the same understanding of value but I did a better job of picking the players that produced it. To that end, the important goal is for categories to be balanced against one another, to reward different approaches and different types of players (hence why I loathe points leagues), and I contend that there’s nothing really wrong with wins from a game mechanics perspective.

    Vote -1 Vote +1

    • atoms says:

      Of course it’s a game. But the point of a game is that it should be, well… gameable, right? The whole point of this post is that, as a fantasy stat, wins are annoying because they are way more like random outcomes than they ought to be. That makes it needlessly more difficult to strategize effectively. Even if you’re good at doing solid player valuation and projection, the variability of the win statistic doesn’t reward that.

      A well-designed game should reward skillful play. Replacing wins with a less random, more projectable stat should make for a better game.

      Vote -1 Vote +1

  22. Bounty says:

    I think QS should mean
    6+ innings and ERA for this game < 4.6 (or some other arbitrary number)

    Vote -1 Vote +1

  23. Jonathan says:

    I like QS combined with IP and K/9, but with no innings cap, as a nice balance. Thus, you can try to rack up your IP by throwing out any and all starters, but you might seriously harm yourself in K/9, ERA, and WHIP. This balances out nicely, in my opinion.

    Vote -1 Vote +1

  24. Dwezilwoffa says:

    Hey I had an idea of innings pitched minus runs allowed divided by starts.
    IP – RA Would have to handle spreadsheets but it makes sense.
    9 IN – 3 RA/1 = 6pts
    6 IN – 3 RA/1 = 3pts

    Not an arbitrary stat, but a counting one.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>