What about Saves – and These Alternatives?

I have enjoyed many of the discussions from the last couple of weeks in response to my invitations into innovation. I’ve expressed my displeasure with one of fantasy baseball’s most notorious categories. Then, because the feedback was so good, I asked what else those who play our game would like to see done differently as far as categories are concerned. The reaction has implored me to collect the ideas and sift through them to see what we can come up with and then take the conversation to another phase. I have also preliminarily developed a couple of suggestions that are intriguing, at least to me.

I realized after I’d posted the second one that I’d left out something, however. I’d wanted to ask what opinionated, creative rotisserie and head-to-head baseball game players thought of saves. Aw hell, let’s include the indifferent and unimaginative ones, too! OK, some of your reactions addressed that topic. And if you really expressed your feelings on saves, then I might have to censor all the curse words.

We have talked about some simple substitutions. SV+HLD and SV-BS and categories of that sort are in our periphery. Specifically, however, I’m curious about what readers thought of possible replacements for the saves category about which I’ve read in my travels but which haven’t caught on. And we really couldn’t have expected them to catch on.

The first comes from former FanGraphs writer Steve Slowinski: “Shutdowns, Meltdowns and Saves.” He later pointed out that the dialog from which they emerged had origins elsewhere, as did the creation of those first two stats. His goal didn’t seem to have much to do with our game, but the idea is really interesting. The “new” pair has been around for several years. It’s derived from Win Probability Added. It’s a pretty simple, objective entry. It’s about the quality of the relief performance. It could in fact help use to derive a counterpart to the quality start or Pure Quality Start, such as the quality relief appearance.

The second comes from Todd Zola, one of my longtime industry colleagues at Mastersball. Todd called it “KNIP: the New Saves.” It’s K/9+IP (innings pitched), in essence. Some commenters on my two blogs have suggested innings pitched for the purpose of a replacement for W, but I haven’t seen this idea for relievers. As Todd notes in his piece from 2012 (keep that in mind when you examine his lists), “the telltale aspect of the K/9+IP list is the absence of a reliever until Craig Kimbrel at #36. Even more relevant is the subsequent order of relievers, which is a far better measure of the player’s skills and contributions.” Note that he removed starting pitchers from the lists displayed, but he explains why earlier in the post.

I think they’re interesting – good talking points. I hope to offer another idea, one that incorporates the use of inherited runners stranded, which I think is an invaluable if imprecise measure of a reliever’s ability. We have a couple of interesting, distinguished concepts here already. How much do they appeal to you?




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.


8 Responses to “What about Saves – and These Alternatives?”

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

    Nicholas,

    I believe there are tables showing expected runs depending on the number of runners on base and the number of outs. While not perfect, this table would give us the ability to compare ‘Runs Saved against Expectation,” whether as a cumulative number or a ratio. If you could somehow factor in a value for closeness of game, I think you might have a number that measures overall effectiveness. Might I suggest adding or subtracting an eighth depending on preference from a scale that includes: +1, tied, +2, -1, +3, +4+, -2, -3+.

    We all know that saves often come earlier than in the ninth when some closer comes in to mow down the seventh-through-ninth hitters with a three-run lead, after his team bumped the lead to three in the top of the inning with a pair of runs. And in the eighth, some setup man REALLY saved the game by coming in and escaping a bases filled jam with no out. Isn’t giving the closer credit for a half-run RSAE before closeness factors and the setup guy almost two RSAE, a better evalutation of their performances? And maybe the save rules could be changed to give the highest RSAE, in the case of a win by less than four runs, the save.

    And if RSAE or Closeness-Adjusted RSAE doesn’t do it for you, how about comparing performance as a ratio against expected on-base percentage of the hitters faced? Just spit-balling ideas for you to do all the work [G]. Thanks for expressing a willingness to do the work and to listen to starter ideas.

    Very Much Appreciated. GM

    Vote -1 Vote +1

  2. David says:

    What our league did, and it works pretty well, is the following: (1) Eliminate the distinction between SP and RP. There are only P. (2) Eliminate saves (no holds either) (3) Use K/9 and HR/9 Allowed. (4) Daily lineups … it won’t work without daily lineups.

    The winning pitching strategy is to carry 4-5 solid ratio RPs that sit in your lineup every day, and rotate SPs through as they actually pitch.

    Vote -1 Vote +1

  3. Melk Was a Bad Choice says:

    Count Saves as Holds.

    Vote -1 Vote +1

  4. Kris says:

    This is the same dilemma that I deal with in hockey. Is the point of fantasy to have a fun game or is the point of fantasy to actively reflect a players value? I don’t believe they’re mutually exclusive, but I tend to have more fun with guys that I know rather than very serious baseball fans. If someone doesn’t have time for hours of prep, they should be able to fire open whatever website they choose and get a decent set of rankings. At least a competitive set of rankings.

    With that said, I’m all for adding quality starts to devalue saves while ensuring there’s an innings limit. It basically ensures that relievers will fall outside the top-75ish, top-100ish based on Z-scores, while also making sure that you can’t just stream it to death.

    I think if you start eliminating saves, you open the window to eliminate a handful of other categories and you’re left settling on a league that only half the guys understand. In a competitive league, you’re obviously just going to points.

    Vote -1 Vote +1

  5. Nodlew says:

    I play in a Roto Yahoo league with Net Saves and Holds. In there, set up men can get credited with a -NS as early as in the 7th inn sometimes. I don’t like it but it is what it is.

    Vote -1 Vote +1

  6. Larry says:

    I recently sent this out to the owners in our fantasy league. We are an 18-team league currently in our 34th year of existence. (We began playing the game in 1981.) For clarity, I have posted the missive in its entirety. I would be interested in any thoughts that any of you might have.

    A New Approach to Fantasy Scoring

    As most of you are aware, the Capital League scoring system has evolved over the years. We started out doing the Roto thing but after a number of years switched over to a points-based system. At the time this was pretty radical and quite a departure from the mainstream fantasy world. We forged ahead anyway and as time passed, fine-tuned that system. We added Total Bases (in place of Home Runs) and HBP to the batter stats, for example, and Holds and Quality Starts to the pitcher scoring. All good moves at the time, but times change and that system, in my humble opinion, is starting to show its age.

    So what’s wrong with the current Capital League scoring system? A couple of things. First of all, with regard to Batter scoring, our current system is at-bat based. The more at-bats a player gets, the more fantasy points he will score, on average. This not only gives those players who bat higher in the lineup an advantage but it also fails to penalize those low on-base-percentage batters who consistently make a higher percentage of outs no matter where they hit in the order.

    In baseball, outs are the clock. A team has 27 outs to score as many runs as possible and making an out diminishes a team’s chances of scoring runs and winning games. The Roto system addresses this reality by using either Batting Average or On Base Percentage as one of its scoring categories. Our system, however, completely ignores the negative impact of making an out.

    Secondly, pitcher scoring in our system is based largely on categories that are game-situation dependent. Holds, Saves, Quality Starts, Wins and even Earned Runs are not based solely on player performance. They are situation-based categories that are loosely tied to performance and using them for fantasy scoring purposes will often times reward pitchers with fantasy points even when they pitch poorly. Holds are the most obvious example of this but Saves are only slightly less nonsensical and the other pitching categories mentioned are also based, in large part, on the circumstances of the game. This makes for a sloppy method of scoring and it makes luck, good and bad, a major contributing factor in how pitcher performance is scored.

    With these issues in mind, I started searching about for other possible methods of scoring player performance. I began working with Linear Weights (LW), a method of rating a player’s offensive contributions by assigning numerical values, or weights, to the outcomes of a batter’s plate appearances. Using Linear Weights as a foundation, I fashioned a fantasy scoring system that I’ll describe in the paragraphs to follow. But before we jump into that, I think a little background might help to lend this idea some credibility.

    While the game of baseball itself has not really changed all that much over the years, the analysis of what is actually going on between the lines certainly has. The great thing about baseball, of course, is that there is so much more to the game than just watching the games. Not to say that the games aren’t fun to watch, they are, but baseball has an additional level of enjoyment that no other sport has: statistics, and consequently, statistical analysis. And when you stop and think about it, there is a very good chance that Fantasy Baseball has been a major motivating factor in the increased interest in the analysis of the statistics of the game.

    For whatever reason, there have been a number of books published in the last ten years that dig deeper into the statistical aspect of the game. Baseball Between The Numbers, The Book, and The Hidden Game of Baseball are pretty much recognized as the cream of this crop. One of the first things you learn from reading books like these is that many of the old-school statistical tools that we grew up with are not all that beneficial when it comes to realistic evaluation of player performance.

    Wins, for example, while credited to a pitcher, are really a team stat. This is because a Win is not based solely on pitcher performance but also on the number of runs that a pitcher’s team puts on the board. The same thing applies to Runs Scored and RBI; they too are really team stats. The number of runs that a player scores will depend on the abilities of those batting behind him and RBI are a product of having runners on base to drive in. The new statistical approach promoted by these books is to remove this sort of statistical “noise” from the analysis of player performance.

    This new line of thinking, referred to as sabermetrics, becomes important to the fantasy game because eliminating this statistical noise from a fantasy scoring system would make the scoring much more accurate. A sabermetrical approach to fantasy scoring would have the affect of taking a lot of the “luck” out of the scoring, making it less like roulette and more like real baseball.

    One of the tools first developed to give baseball fans a better method of evaluating player performance was the previously mentioned Linear Weights system. Created by a guy named Pete Palmer (The Hidden Game of Baseball), the LW system places a “runs created” value on the results of a batter’s plate appearances. After crunching a whole lot of numbers, Palmer determined that a Home Run would result in approximately 1.4 runs being scored. A triple would, on average, produce 1 run; a double, .8; a single .46, and a walk, .33. A stolen base was worth approximately .25 runs. Also, since making an out diminishes a team’s scoring chances, the LW system deducted points for outs, both at the plate and on the base paths.

    This is very interesting stuff and I think you can easily see why the idea of using Linear Weights as the basis of a fantasy scoring system came to mind. The problem with Linear Weights, however, is that it is an analytical tool for batter production and, as you know, fantasy baseball includes pitchers as well. The challenge was to create a corresponding weight-based system for scoring pitcher performance, while at the same time maintaining the balance between pitcher and batter scoring that we have come to expect.

    While struggling with this problem, I ran across a Runs Created system developed by a guy named Paul Johnson. Unlike Palmer, Johnson used whole numbers in his system (weights of 9, 7, 5, 3, 2, 1) instead of the decimals that Palmer used. I immediately jumped on Johnson’s bandwagon. Not only were whole numbers a lot easier to work with, but this guy had already done the research regarding the validity of using those numbers to compare player performance. I could use them with confidence. So here’s what I came up with.

    The scoring for batters under this system is similar to our current system. I simply substituted the weights developed by Paul Johnson, as indicated above. Stolen Bases and Caught Stealing are included in this new system but there are no points awarded for Runs or RBI. The new system also penalizes a batter for making an out.

    The scoring for pitchers, on the other hand, gets a major makeover. Under this system, pitcher scoring is based solely on the outcome of the pitcher’s duel with each batter. Mano-a-mano. Wins, Quality Starts, Holds, Saves and Net Scoreless Innings are no longer a component of pitcher scoring. Simply put, this scoring system rewards pitchers for getting batters out and penalizes them for allowing batters to reach base. This system uses the same weights as those used to reward batter performance, only now reversed as a negative to the pitcher.

    On the positive side, a pitcher is credited with two points for each out recorded, or six points for an inning pitched. Pitchers are awarded additional points for keeping the ball out of play, that is, for striking batters out. (Runs can score on fly outs or ground outs. A poor defense can turn balls-in-play into hits or errors. A strikeout keeps the ball out of play and slams the door on all of these negative scenarios. Hence, additional points awarded.)

    A relief pitcher works about a third as many innings as a starter and this creates a problem when it comes to rewarding them for the work they do. The truth of the matter is, the innings a relief pitcher typically pitches are almost always critical to the outcome of the game. This system gives relief pitchers credit for just appearing in these more intense, high pressure situations by awarding them the equivalent of an extra out (2 points) just for appearing in the game.

    So that’s the basic idea. Scoring for both pitchers and batters under this system would focus almost exclusively on the outcome of each at-bat. This would distill the game down to its basic component, the batter versus the pitcher. Except for stolen bases, caught stealing and relief appearances, all scoring would be based on the outcome of the Batter/Pitcher matchup. Here’s a look at the scoring.

    Batter Scoring:
    Out Created (AB – H) or (GDP) (-.5); Single (+3); Double (+5); Triple (+7); Home Run (+9); Walk or HBP (+2); Stolen Base (+1); Caught Stealing (-1).

    Pitcher Scoring:
    Inning Pitched (+6); Relief Appearance (+2); Strikeout (+1.5); Single Allowed (-3); Double Allowed (-5); Triple Allowed (-7); Home Run Allowed (-9); Bases On Balls (-2); Hit Batter (-2).

    Now just to be clear, I am not suggesting that the Capital League switch over to this scoring system. This discussion is more about fantasy scoring in general and I’m just throwing out ideas here. My point is that whether one is playing in a Roto or a Points-based league, the scoring systems currently being used leave a lot to be desired. I think the whole world of Fantasy Baseball is starting to show its age and is in need of a scoring system remodel.

    This becomes even more obvious as statistical analysis continues to evolve. New concepts such as wOBA and wRC+, and for pitchers, FIP, xFIP and SIERA are now the analytical tools of choice. The latest is a one-size-fits-all designation called WAR (Wins Above Replacement). WAR is context, league and park neutral. WAR combines both batter and pitcher performance (by way of runs created and runs prevented), as well as defense, into one number. In my humble opinion, these exciting new tools of player analysis have left fantasy scoring systems looking awfully antiquated and I think the fantasy game would be foolish not to incorporate some of these progressive ideas into the scoring. I only wish that someone with the ability to create an analytical tool such as WAR would turn their attention to fantasy baseball and design the ultimate scoring system. What you’ve just read is simply a layman’s attempt at doing that.

    Vote -1 Vote +1

  7. novaether says:

    I like SD-MD, or at the very just SD, to be the eventual replacement. People would be scared at first since very few people could tell you off the top of their head exactly which scenarios are possible SD scenarios, but I think people would get used to it pretty quickly. I imagine a table of SDs would contain all save situations, mop up situations, and then protecting small leads in the last few innings. Does anybody have a table of SDs or at least a more rubust table than this?

    http://www.tangotiger.net/welist.html

    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>