By In Stuff

If I were baseball stats king …

So, I’m getting a lot of questions from Brilliant Readers about what I would do to make mainstream baseball stats less judgmental. Well, OK, they’re not all framed exactly like that. Some of them start with something along the lines of “OK, genius, what would you do if …” or “Yeah, baseball statistics have been around for 100 years, how are YOU going to improve them?” But I think the point remains the same.

So what I’m doing here is working only with mainstream statistics — the sort you will hear talked about regularly on radio and television broadcasts. I realize that none of my suggestions will ever be embraced, and that’s almost certainly for the best. But, well, since you asked, if I were named baseball stats king, I would do some of these things:

Errors: II would immediately stop using errors the way they’re being used. As it stands now, for batting average and on-base percentage, reached-on-error is currently counted as an out. I loathe that. There was no out recorded. You cannot just make up outs because you believe they should have happened. You cannot credit (or debit) a batter with an out if he did not make an out. It’s terrible bookkeeping.

I personally dislike the error so much, I wouldn’t care if it went the way of the dodo. I think it’s pointless. And I think it does a terrible job of doing the one thing it’s supposed to do: Penalize poor defense. I saw another example of this Wednesday in the Tampa Bay-Texas game. In extra innings, Texas had the super-speedy Elvis Andrus on first base with two outs when Adrian Beltre singled to right. Tampa Bay’s right fielder Wil Myers — whose average is back at .300 again because he’s been killing the ball lately — did two terrible defensive things. First he took a bizarre, “I shoulda made a left turn in Albuquerque” angle to the ball and so took about three months to get to it. Second, he seemed throughly unaware that Andrus, who had gotten a great jump, was a threat to score from first on the play. It was like the idea never even occurred to him. So he kind of casually threw it to the cutoff man. Andrus did not just score, he scored EASILY. It was a disastrous defensive play. But, guess what? No error.*

*To Myers’ credit, he immediately realized his blunders and when the Rays scored the tying run in the bottom of the inning, Myers looked up at the heavens (or at the roof of Tropicana Field) and mouthed: “Thank God!”

But I realize that the error does serve some purpose for many. It’s familiar. It’s a part of baseball’s language. It connects our game to the past. It’s fairly simple. So here’s what I would do to placate the error-lovers out but still make some sense of the accounting: We still count errors. We count them for the defender. And if a player reaches base on an error it would count as a ROE (Reached On Error) and not as a hit.

BUT … for batting average, it would NOT be recorded as an out. That’s just lunacy. It would count the way a walk counts — no at-bat. And for on-base percentage, it would credit the batter for getting on base because, um, the batter got on base.

As far as errors that allow runners to gain extra bases, I’ll leave those for another time.

ERA: So then you might ask — without errors what would I do with Earned Run Average? ERA is actually a funny thing, I hear people lambaste various forms of FIP — Fielding Independent Pitching stats — because they say you can’t separate pitching from fielding. But really ERA was the first form of FIP. The idea was to try and isolate what the pitcher had control over (earned runs) and what he does not (unearned runs). I think of ERA as a crude but noble statistic and, again, like errors, I realize that ERA connects us to baseball history.

But, personally, I think separating ERA from simple RA is pointless and distracting and often misleading. An unearned run counts just as much as an earned run. I personally would prefer to know a pitchers run average. So that’s what I would do — I would make RA the statistic of choice rather than ERA. Yes, I realize it’s not quite as romantic to say that Bob Gibson had a 1.45 RA in 1968 (rather than his famous 1.12 ERA) or that this year Clayton Kershaw has a 2.22 RA (rather than his 1.94 ERA). Yes I realize that we have grown used to the rhythms of ERA — with a sub-3.00 ERA being good, a sub-2.00 ERA being outstanding, a 5.00-plus ERA being dreadful. I still think Run Average tells us more.

I wish RA was more accessible.

Sacrifice flies and sacrifice hits: I would do two things with sacrifices. One, I would count them as outs in batting average and on-base percentage. It’s like the opposite of errors. You can’t just pretend an out wasn’t recorded because you admire the player’s benevolence.

And, anyway, not counting a sacrifice as an out in individual statistics goes against the WHOLE SPIRIT OF SACRIFICE. Sure, I know, sacrifice is supposedly referring to the team sacrificing an out for something else. But I’m sorry, we hear all the time about how unselfish these bunters and sac-fliers are — I think they should make PERSONAL sacrifices to their batting average. I mean, sacrifice is in the very title of what they’re doing.

So that’s the first thing: Sacrifices count as outs in batting average on on base percentage. Elvis Andrus is hitting .271 this year? No he isn’t. He’s hitting .262 when you count in his 13 sacrifice hits and six sacrifice flies.

But, the second thing is: I would not only count sacrifices, I would expand their definition. I think anything that accomplishes what sacrifices do should count as sacrifices. A sac bunt moves a runner from first to second, right? Well, to me any batted ball that does that should count as a sacrifice. I’m not interested in guessing at the player’s motivation. If he check swings a runner to second, if he high chops a runner to second, if he squibs or fists or jam-shots a runner to second, I’d count it as a sacrifice hit. Remember, sacrifice hits count as an out. But your contribution to the team is noted for the record.

Sac flies … same thing. Any out-producing RBI would count as a sac fly for me — we could change the “Sacrifice fly” to “Sacrifice score” or something. I’ve always thought it illogical that a fly ball to the outfield that scores a runner from third is considered a sacrifice but a ground ball that scores a runner when the infield is back does not. It’s the same thing. Any time you score a runner with an out, you get a sacrifice score.

No RBI on double play: Loathe the rule as it is now where batters do not get RBIs when they hit into a double play and the run scores. Did he drive in the run? Yes. That’s the only thing this stat is supposed to be counting. I think the RBI should mind its own business and let double play concerns be dealt with by more capable stats.

Defensive indifference: Loathe it. Get rid of it. If a runner steals a base, he steals a base. I don’t care if the team didn’t try to throw him out. That’s their business. A stolen base is a stolen base.

Reaching on a dropped third strike: I would treat this the same way I would treat errors. No out was recorded so no out should be recorded in batting average. The pitcher should not get credit for the strikeout either — he didn’t strike the batter out, did he? The batter reached base so he should be credited of reaching base in on-base percentage.

Pitcher wins: Ah, our Brian Kenny section. Much like errors and ERA and so on, I realize that pitcher wins has a connection to baseball history and it has come to represent a common language. I find myself using wins and losses as shorthand to get points across and I think the win is ridiculous. So I understand why people might think Brian’s “Kill The Win” is a bit harsh. But I also find wins to be useless when actually judging a pitcher’s performance, and it is nails-against-a-chalkboard at this point to hear people who should know better try to inject some sort of magical quality to winning pitchers.

Here’s one thing I would do to improve the win — I would simplify the statistic. I’d get rid of all the goofy pitcher rules and make it simple. If you start a game and the team wins, you get the win. If you start a game and the team loses, you get the loss. That’s it. No more reliever wins. No more no-decisions. The starter — much like an NFL quarterback — would have a record based precisely on how the team did.

Would this make the win any more credible as a statistic? Of course not. In some ways it will make the stat LESS credible for a million reasons I’m sure you’ve already come up with — one example being that you will have pitchers win extra inning games when they were out in the third inning. But we’re already dealing with a silly statistic. What I like about using team wins is that, unlike pitcher wins, it gives me actual information I can use.

Take Max Scherzer. He is 19-3. OK, great, what does that tell me? Almost nothing. Scherzer has made 30 starts. In the 22 games we decided to count, his team won 19 of them. That’s what it tells me. That’s almost nothing.

Now, if you tell me that the Tigers have gone 23-7 in his 30 starts, OK, now that’s actual information. I might not use that information the way you would use it — I might not use it at all — but it isn’t bogged down by any judgments. If I say that the Tigers have won 23 of the 30 starts Max Scherzer has made, that’s unfiltered truth. Meanwhile, Max Scherzer’s 19-3 record is just what this silly statistic tells me.

Let me add: Team wins is a very difficult stat to find on the Internet. Best I can tell, the best way to find it is to dig deep into Baseball Reference. And you can’t really sort that information — I’d like to be able to do that. I wish true pitcher wins was more readily accessible.

83 Responses to If I were baseball stats king …

  1. Richard says:

    RA sounds a little spare. Maybe you could call it PRA for Pitcher Run Average. Or DRA for Defensive Run Average. Either way, a run is a run.

  2. Jason says:

    What about the save? Merge it with holds would be my suggestion.

  3. Matt says:

    I’m OK separating out Defensive Indifference from SB, just like I’m OK separating out IBB from BB. But at least keep track of it!

  4. What about fielders choice, should that improve obp? I got on base after all, just like the logic used with an error.

    • It’s not the same. In a fielders choice, an out is recorded. In the error scenario, no out is recorded. Follow the outs.

    • Tonus says:

      There was an out recorded, so I can understand why it would count against the batter, but it’s still a good question. After all, if a batter hits a single and the runner on first is thrown out at third, the batter is still credited with a single. That one could be tricky, or there’s something I’m missing.

      As for stats, I’d do away with the save or change the rules so that a save is credited when a reliever enters the ninth inning with a 1-run lead and runners on base and does not allow a run. Or perhaps have it reflect higher-leverage situations. If one reliever escapes a jam in the eighth and another reliever finishes a 1-2-3 ninth, maybe the guy who pitched the eighth inning deserves the save? But that makes it subjective. Kill it.

    • J Hench says:

      There’s also the fielder’s choice where no out is recorded. Alex Presley hit one tonight in the Twins-A’s game. Runner on third, Presley hits a chopper to second, Sogard throws home but just a moment too late to get Parmelee at the plate. Twins get a run, Presley gets on 1B, no out recorded, but it counts against Presley’s batting average (and OBP).

      Which, by the way, I don’t think the issue that is bugging Joe about errors is the errors; it is in terms of how batting average is calculated. This is similar to the complications with pitcher wins. The concept behind batting average is simple: let’s figure out what percentage of the time a batter gets a hit out of the opportunities he has to get a hit. Totally makes sense. But then you start figuring what counts as an opportunity — a walk can’t count as an opportunity. He didn’t have a chance to get a hit on a walk. But a ground ball to the second baseman; that’s an opportunity, and he didn’t get a hit. Would have been an easy out, if the second baseman wanted to make it an out. Sac bunt, well, he wasn’t trying to get a hit, so that’s not an opportunity. And so on. There’s a logic to it as long as you understand what batting average is trying to measure but it makes it quite complicated. I do think times on base should count towards OBP, that being a calculation … times on base.

    • Schere says:

      not all FCs result in outs, of course.

    • Schere says:

      not all FCs result in outs, of course.

    • JRoth says:

      Don’t confuse the issue wit dirty facts, Schere. Joe is in the realm of the Empyrean.

  5. Ross Holden says:

    What about no error because you can’t assume a double play? I agree errors are lame, but if we still count them for the defender, it should count as one if a first baseman drops a ball that would complete a DP.

    • Matt Janik says:

      Actually, the instance you cited is an example where you CAN charge an error on an attempted double play. First baseman dropping an on-target throw is an error, regardless of if an out was recorded at second (or third) beforehand. The not assuming the double play has to do with the throw by the 2B/SS/3B after already putting out a runner; wild throw there can’t be charged with an error (unless a runner takes an extra base because of it).

      Not that the rule makes any MORE sense the way it’s written (if you throw the ball away when it should’ve been an out, that should be an error, no?), but the example you’ve chosen actually can (and should) be charged an error.

    • Ross Holden says:

      I see what you mean. Thanks for the clarification. I agree, it is still odd.

    • Rob Smith says:

      Well, the context on this play is that the middle infielder has a runner bearing down on him with intent to hit him and disrupt the throw. Nonetheless, the middle infielder is expected to make an awkward throw, while avoiding the runner, and complete the DP. The rule is just acknowledging that completing a DP is difficult and will often result in an off balance throw as contact is made by the runner. So, realistically the middle infielder is trying to make a difficult play in the same way that they would making a throw from the deep hole. In the case of a the play in the deep hole, the shortstop is not reasonably expected to complete the play, so a poor throw will not be given an error… unless, as with the DP, the runner takes another base. That too, I think is fair. There is an element of decision making involved where making a throw to complete an almost impossible play might not be the best play. So, I think the rule is just fine. The team got an out, failed to get the more difficult second out and the stat just acknowledges the difficulty of getting that second out & doesn’t punish the fielder for trying.

  6. Tangotiger says:

    We’ve been calling it RA9.


    In hockey, they still give an assist to the guy who gets an empty net goal, no matter how pointless the pass was. And come to think of it, how pointless the goal was. That’s the ultimate in defensive indifference: we’ll take out the goalie!

    And yet it all counts. They track it as a separate category, but it still gets rolled into goals scored.

    • nscadu 9 says:

      Yes, and hockey goes out of its way to credit the offence. If an own goal occurs the official scorer finds an offensive player to credit. Last player to touch the puck, closest proximity. The player could have nothing to do with the actual goal and still credit is given. Funny how baseball is so tied up in judgement calls discrediting offence.

    • James says:

      In hockey, that’s a sacrifice and it still takes skill to get the puck away from your opponent. With an indifference play, you’re rewarding a player for being on a losing team. It gains his team nothing. There are so many important extra bases in baseball which we need to get accounted before we worry about defensive indifference.

  7. Butch says:

    One I’d change: Simply change the name of quality start to winnable start. People get hung up on the definition of “quality.” How can a 4.50 ERA be “quality”? OK, it’s not quality. But if a pitcher gives up 3 runs and gets you to the 7th inning, is it his fault the team lost? Do you look at the box score the next morning and blame the starting pitcher for a 3-2 loss?

    Winnable start. Try it.

  8. I have been wondering for several years why no one simply counts the bases gained and lost by each player. The rules would be along these lines:

    1. Hitters are credited with the number of bases gained that they achieve themselves, plus up to the same number of bases gained for each runner who was already on base. That includes walks, HBP, reaching on an error, etc. A leadoff single is worth 1 bases gained; a grand slam is worth 10 bases gained.
    2. Runners are credited with the number of bases gained that they achieve themselves (stolen, defensive indifference, wild pitch, passed ball, errant pickoff throw, etc.), as well as any extra bases gained that they take on a hit (going first to third or scoring from second on a single, scoring from first on a double).
    3. Those who make outs on the bases (picked off, caught stealing, thrown out trying to take an extra base, etc.) are docked all of the bases that they had gained up to that point. For example, being tagged out at home costs 3 bases. In addition …
    4. When runners are left on base at the end of an inning, the bases that they had gained are docked from those who made the three outs, starting with the last out. So if there is a man on second when the final out is made, the last two out-makers lose 1 base each; if the bases are loaded, each out-maker loses 2 bases.
    5. Hitters would be credited with bases gained by sacrifice hits/flies, including “unintentional” sacrifices like Joe discusses in this post, as well as fly balls on which a player advances from first to second or from second to third. However, they would still potentially be penalized for any runner left on base at the end of the inning.

    The beauty of this system is that the number of runs scored by the team will always equal exactly one-fourth of its total net bases. Therefore, credit for those runs could be assigned to individual players in the same fashion; no more need for “advanced” statistics like runs created or linear weights, as fun as they are. Likewise, pitchers and fielders could perhaps be measured by net bases allowed. Every player’s offensive and defensive net bases could be reported in the box score and totaled over the course of the season.

    • Rob Smith says:

      I was with you right up until “I’ve been wondering”.

    • indyralph says:

      The simple answer is that your system has too much context dependent noise. A HR is a HR. How many people were on base has nothing to do with the person who hit it. Four singles does not equal one home run, etc. wRC is really not much more “advanced” than your system, and it does a better job.

    • invitro says:

      Basic RC is about 100 times simpler and does a much better job.

    • Which is more valuable, a home run with the bases empty or a grand slam? a leadoff single, or one with men on base? Should they really be treated as completely indistinguishable when evaluating a hitter? Are you sure that four leadoff singles are not at least roughly equivalent to a leadoff home run? While “clutch” hitting is certainly overrated, context is not completely irrelevant or merely “noise.” The idea here is to assign credit fairly for the runs that actually score, as well as the runs that could have scored if the inning had gone differently.

      Run production is a matter of some people getting on base, and then other people moving them around. Simple runs created–which I happen to like a lot–uses OBP for the first component and TB for the second. [In fact, my favorite “mainstream” advanced baseball statistic is simple runs created per out, since it can be calculated directly from the three “slash” numbers as OBP*SLG/(1-AVG).] BaseRuns uses times on base for the first component and an estimate of the percentage of runners who score for the second. Even OPS uses OBP for the first component and SLG for the second, but adds them rather than multiplying–a clear mathematical error (different denominators) that nevertheless produces a useful result.

      What I am suggesting here is another approach in accordance with the same basic theory, but one that reflects what actually happens, rather than an amalgamation of cumulative outcomes devoid of contextual information. That said, I have no idea what the numbers would actually look like over the course of an entire season. I would love to find out, but I do not have the data or the time to do the work myself.

    • invitro says:

      “I do not have the data or the time to do the work myself”

      I think Retrosheet or Baseball Info Solutions should have the data. I hope you’re not asking others to do your work for you :). *You* need to demonstrate that your statistic is better than what is currently used.

      OPS would be fine if it weighted OBP the appropriate amount (twice?) with respect to SLG.

      I use the Clutch column from b-r for awards, and hope that I am correctly understanding what it measures.

    • Tangotiger says:

      What you want is being done at

      I prefer RE24 on Fangraphs or Baseball Reference. The main difference between what you propose and RE24 is that RE24 is denominated in runs.

    • I am not asking anyone to do anything, and I am not arguing that what I proposed here would be “better” than other options. I just think that it would be interesting–another tool in the box.

      The only way to correct the mathematical error inherent in OPS is to adjust SLG to include all of the same terms in the denominator as OBP, and also presumably add some or all of them to the numerator. The interesting thing is that OPS+ already correlates fairly well with runs created per game (25-27 outs depending on what is included) divided by league average runs per game.

      I appreciate the heads-up about, but after taking a quick look, I do not see anything there that matches up with what I have in mind. You cannot determine the number of bases to credit to any player (using my rules above) from his cumulative statistics alone; you would have to go through the play-by-play data.

      Of course, the same is true for RE24, which is likewise context-dependent. My misgiving there is depending on historical averages for run expectancy in each of the 24 runner-out states.

    • invitro says:

      “The only way to correct the mathematical error inherent in OPS is to adjust SLG to include all of the same terms in the denominator as OBP, and also presumably add some or all of them to the numerator.”

      This is false. The point is to have something that correlates with runs scored. From what I’ve read (and done, to a small extent), runs scored correlates more closely to about 1.8 OBP + SLG than to OPS. Or OBP and SLG can be multiplied, as in RC.

    • That may be your point, but it is not mine. My point is that it is never mathematically correct to add two numbers that are calculated using different equations in the denominator. OBP = (H+BB+HBP) / (AB+BB+HBP+SF) and SLG = TB / AB. Only if SLG is adjusted to something like (TB+BB+HBP) / (AB+BB+HBP+SF) would adding it to OBP be mathematically correct.

      Whether the result correlates with runs scored–i.e., the practical usefulness of the statistic, rather than its mathematical correctness–is a whole different issue. As I noted previously, OPS+ actually seems to exhibit that kind of correlation, but I personally prefer simple runs created per out = OBP*SLG/(1-AVG). Multiply this by the league average (AB-H)/G–typically about 25.5–to get runs created per game for comparison with league average runs per game.

    • By the way, since OPS+ is the non-dimensional result of adding the ratios of the player’s OBP and SLG to the corresponding park-adjusted league average values and then subtracting one, it is mathematically correct, unlike OPS itself.

  9. But, by your logic- which I like, by the way- a run was still batted in on a double play ground ball. It’s what happened, so shouldn’t it count all the same as other rbis? That seems to be adding the kind of exception you loathe, like an error not counting as a time on base.

    • J Hench says:

      I think that’s exactly what Joe was saying–he loathes that they don’t award an RBI on a double play. That’s what he means by let double plays be handled by more capable stats. But I had to read it twice to figure out which side he was on, too.

    • Rob Smith says:

      I don’t have a problem with not awarding an RBI on a double play. Yeah, technically he hit the ball and a run scored…. on the other hand, and this is what Joe rants about all the time, two outs were created, and outs are precious, right? You only get 27 of them & one AB just took 7.5% of the alotted outs for the entire game. Besides a triple play, that’s the worst outcome possible in an at bat. So, the idea of giving any positive credit (RBIs) for creating two outs also does not make a lot of sense. At least with the error example (adding it to OBP) you can say that faster runners force fielders to rush and make more errors, or the hitter hit the ball hard enough that it made the play difficult, so there’s some possible influence there. But I’m perfectly OK with not giving a RBI when there is a DP. To me, adding an RBI in that situation just makes the stat more meaningless.

  10. Tom says:

    Imagine that it’s the last day of the season and a player on an opposing team is one stolen base away from breaking a cherished stolen base record held by a former member of your team. You decide to refuse to try to throw him out every time he steals. Result? He steals four bases but doesn’t set the record because none of them count due to Defensive Indifference. Absurd? Yes. Joe is right.

    • Martin F. says:

      Defensive indifference doesn’t work that way. It’s not about if a team tries to throw a runner out, it is if a team doesn’t try to throw the runner out in a situation in which it makes no difference. They don’t award DI in the 4th inning. They don’t award DI if it’s the tying run, even if no attempt was made to throw the runner out.

    • Ian R. says:

      In addition, there’s a rule specifically instructing the official scorer in the situation you described. To wit:

      “The official scorer may conclude that the defensive team is impermissibly trying to deny a runner credit for a stolen base if, for example, the defensive team fails to defend the advance of a runner approaching a league or career record or a league statistical title.”

      It’s a judgment call, but I doubt any scorer would call such a situation defensive indifference, and I imagine the Commissioner’s office would overrule the scorer if he did.

    • Rob Smith says:

      Thanks Ian for correcting the idea that defensive indifference can’t be called in the 4th inning. That’s simply not the rule. It’s a complete judgement call that often, but not always, takes place in a later inning where the defensive team, often with runners on 1st and 3rd, makes the decision that (like with having the infield in)that even more run scored just puts the final nail in the coffin. This can happen earlier in the game, just like having the infield in, although it certainly happens that way less often. So, they can’t risk a throw to second on a stealing runner because of the (fairly high) possibility of the runner on 3rd scoring. But I do think defenisve indifference is a dumb rule. If you steal second, you steal second. It wasn’t some sort of manipulation of physics that got the runner to second. The runner took a calculated risk and stole second. Rarely does a runner know ahead of time for certain that the catcher won’t try to throw them out.

  11. Martin F. says:

    I’m not sure that RA does anything except change the numbers used. Counting the RA doesn’t tell me anything that ERA doesn’t, except it makes it less specific, and makes the numbers larger. What’s the point? If you use 2 statistics, ERA/RA, then at least there can be a shown comparison and you can get some idea of how good a fielding team there is behind the pitcher.

    1.45 vs. 1.12

    That tells me nothing other than Gibson had a great year. It’s only if we delve into % of runners who reached on error who scored, that we get some useful information. It becomes analogous to Inherited Runners who Score %, which leads to another beef.

    Why is it that the pitcher who leaves the game with the runner on first gets charged the RA? Why isn’t the reliever charged the RA, since he’s the one who allowed the runner to score? If RA is used to show a pitcher allowed a run to score as a result of an error, why not use it to show which pitcher allowed the runner to score? If Wainwright walks a man and then leaves the game, should it be his fault that Mujica blew up like a fireworks display with a couple walks and doubles?

  12. Dinky says:

    While I in general abhor the sacrifice, it is rarely done by the batter without being told to do so by the manager. So penalizing a batter’s statistics for something he is told to do seems wrong. Similarly, on a ground ball with a runner on third and nobody out in the first inning, you will probably drive in that run. But if it’s tied in the ninth, the infield will be in and the ground ball will either be a single or no run will score. A deep fly ball will score the runner no matter the game situation. Thus, batters who are capable of getting sacrifice flies are worth a bit more in likely RBI situations than guys who hit it into the dirt all the time. A guy who can hit it 390 feet to right field won’t get credit for a runner tagging up and moving to third, but a guy who puts down a bunt does get a sacrifice. Now ask yourself: would you rather have a team of guys who can bunt or guys who can hit the ball hard? So why does one type of hitter get the benefit of a sacrifice for advancing a runner but the other hitter doesn’t? And a fly ball to the warning track is a safer runner advance than all but the best placed bunts.

    I agree with errors: if you hit the ball, and the fielders are unable to get you out, you deserve credit for it. If you’re at home, you’re much more likely to have the official scorer credit you with a hit than if you are on the road on the same hard hit ball that isn’t fielded cleanly.

    The real issue is to replace scoring decisions with team value decisions. I am on your side with that.

    • Wilbur says:

      “While I in general abhor the sacrifice, it is rarely done by the batter without being told to do so by the manager. So penalizing a batter’s statistics for something he is told to do seems wrong.”

      A runner caught stealing is usually told to steal by the manager … should his statistics be “penalized” for that?

    • Robert says:

      “Thus, batters who are capable of getting sacrifice flies are worth a bit more in likely RBI situations than guys who hit it into the dirt all the time.”

      Circle me, Joey Gathright.

    • Rob Smith says:

      Wilbur your analogy is not accurate. If a runner is told to steal, he has the possiblity of success. Even with a slower runner, who are rarely asked to steal, the pitch has to be handled by the catcher and thrown accurately to second. That’s not a foregone conclusion. With a slow runner, there is a big element of surprise & the pitch might be a slider in the dirt. Good throw, bad throw or no throw, success = stolen base. With a sacrifice, though there is a very slight possiblity of a hit, the general idea is that you will put all efforts into getting the bunt down with the understanding that you will create an out. It’s not a bunt for a hit and see what happens scenario. It’s trading an out for a base. Obviously the stat was manipulated so as to not punish the batter for being asked to bunt, especially those that are asked to do so frequently. If sacrifices knocked 10 pts off your batting average, as with the example given on Elvis Andrus, players might not be so agreeable about sacrificing. As it is, most players hate it when they are told to sacrifice. Everyone, with the possible exception of pitchers and very poor hitters, wants to swing away.

  13. Alejo says:

    Pretty much agree on everything except wins because what you propose is as silly as what you decry.

  14. Brendan says:

    I like most of these ideas. An interesting (to me) tangent thought: if you were to redesign the box score, how would it look? My idea: columns with plate appearances, total bases, total outs, runs scored.

  15. Mark Daniel says:

    The ROE one seems simple, but it’s not because it just so happens that these days there are fewer errors per game than at any time in baseball history (0.55 errors per game in 2013 in the AL, down from 0.59 last year and 0.69 in 2000). Moreover, the error rate goes up steadily as you look back in time. By decade (AL only):
    2010, .62
    2000, .69
    1990, .73
    1980, .85
    1970, .84
    1960, .85
    1950, .92
    1940, 1.16
    1930, 1.26
    1920, 1.39
    1910, 1.85
    1901, 2.62

    So, if you look at someone like Ty Cobb, who played from 1905-1928 (E rate of 1.77-1.25, respectively), you would expect him to have gotten on base by ROE an awful lot. But they didn’t keep track of ROE back then. So Cobb just gets outs counted. Therefore, if you changed the stats so that ROE counts much like a walk, that would hinder accurate comparisons of modern players to players from the past.
    ROE is already counted in WAR, but I don’t know if ROE is adjusted for the error rate in a given year. I also don’t know how it reconciles the lack of ROE for someone like Cobb.

  16. Wilbur says:

    I’d still like an answer to a question I posed a few threads ago: How do you deal with an error where nothing bad occurs as a direct result – like a dropped pop foul?

  17. cd1515 says:

    the difference between a pitcher and an NFL QB getting credit for the win is that usually 3 other QBs don’t follow the starting QB in the same game.

    pitcher wins should be at least 6 innings though.
    if they were, you’d suddenly see many more starters “lasting” 6 instead of 5.

  18. Robert says:

    This comment has been removed by the author.

  19. Eric H. says:

    Just wanted to point out that sacrifice flies already DO count as outs in OBP, but not in batting average. I’m not sure why.

    • The theory is that the batter is not trying to hit a sacrifice fly, but rather just trying to drive the ball to the outfield (base hit, HR, whatever). Basically, he was trying to get on base with a hit of some sort, didn’t, but his out was productive, so don’t ding him for BA. Contrast that with a sac bunt, where the batter supposedly isn’t really trying to get on base.

      I’m definitely not saying it makes any sense (e.g. bottom of the 9th, winning run on third, batter doesn’t care if it’s a sac fly or HR; or maybe Rickey Henderson was really trying to get on base w/ a bunt that happened to score a run), but that’s the theory behind it.

    • Rob Smith says:

      Yeah, I thought you were wrong, but I looked it up & yep, SF’s are not counted in OBP. Odd, since one would think OBP would be based on ABs, just like BA… but they actually add the SFs back to the ABs. So it was done on purpose for some reason. You could say, well, the player didn’t get on base so it shouldn’t be in OBP… but Sacrifice bunts are still included even though they didn’t get on base. That’s a little crazy. But, I guess that’s the point of this blog. The stats don’t always make sense.

  20. Trent Phloog says:

    I am deeply curious about the “team” wins stat and how the record books would look different using it. Is Cy Young still the record holder? Or someone else who pitched forever, like Nolan Ryan? It would have to be among the career leaders in games started, right?

    Joe, can you provide more info about how to “dig deep into Baseball Reference” to find this stat? I can’t find it at all…

    • BeefMaster says:

      On a player’s page, click the “More Stats” tab next to “Standard Pitching”. On that page, scroll down to “Starting Pitching”, and it’s the “Wtm” and “Ltm” stats.

      Nolan Ryan’s teams were 405-368 in games he started. They don’t have that data for Cy Young, probably because his career was before the Retrosheet era, so they can’t narrow down all the specific games he started.

  21. Richard S. says:

    Here’s a thought…

    Thanks to many fans, statisticians, archivists, Sabermetricians, and the like, we have an immense amount of raw numbers about baseball available to us. Take whatever set of revised / alternate statistics you wish, and recalculate the numbers for an entire season. Or, watch a month’s worth of games (that’s at least 30 games in total) and play “Official Scorer”, assigning errors (of whatever definition), “defensive indifference”, et al. as they happen.

    I’m willing to wager that it won’t make a noticeable difference. Sure, a few things like batting titles might get juggled around a bit, but Walter Johnson, Tom Seaver, Mel Ott, and Frank Robinson will still be awesome.

    • invitro says:

      Don’t you think there is a rather enormous gap between “Walter Johnson, Tom Seaver, Mel Ott, and Frank Robinson will still be awesome” and “won’t make a noticeable difference”?

    • BeefMaster says:

      Yes, the fact that something mostly works means we should never seek to improve it.

      I’m assuming that you typed your post on a computer running Windows 3.1, as well.

  22. This comment has been removed by the author.

  23. This comment has been removed by the author.

  24. Go into the pitcher’s Game Logs, and the very top line is “Team Record in Appearances.” You have to go year-by-year, but it’s really not very hard.

  25. Mike P says:

    If RA became accepted, RA Dickey could join Homer Bailey in having the worst name related to his position in baseball.

    I’d make a double play count as two outs on a Batting average/on base percentage.

  26. DJM says:

    I mentioned this in a comment thread in a recent post, but it might be easier to see here.


    When a pitcher enters a game with the tying or go-ahead run on-base or up to bat, and completes the inning without allowing a run, they receive a save. Each pitcher in a game who meets those requirements may receive a save.

    This does several things. First, it takes the save back to the original point: rewarding relief pitchers for doing their job. Also, it would make save opportunities higher-leverage situations overall, eliminating the now-traditional ninth-inning save in most forms. Finally, it will increase the overall number of saves, devaluing each individual save and making it less likely a team will “pay for saves” for their own sake.

    • adam says:

      Seems reasonable. I prefer to get rid of the save entirely, but that’s a good compromise.

    • nickolai says:

      I hate it. If you did this, my beloved Billy Beane wouldn’t be able to fleece other teams by trading them ‘established closers’ like Billy Koch, Huston Street, Andrew Bailey,…

    • Eric H. says:

      I like this idea better than the current save rule, but it still has some problems. For instance, what about a guy who “saves” the 6th inning, then gives up 6 runs in the 7th? Does he really deserve a save?

      The existing save rule, flawed though it may be, does give you one valuable piece of information: The pitcher’s team won the game, so he pitched at least well enough to “not lose.”

      With your revised save rule, a middle reliever could absolutely blow up on the mound, and still get credited with a save (and a loss).

  27. Underlying the concept of the pitcher win is the notion that the pitcher has the most responsibility for keeping the opposition from scoring. That is probably true, although it made more sense when pitchers routinely pitched complete games. Today, while the pitcher certainly may contribute to the win more than any individual, he certainly is not solely responsible for the game score. It is always a team win, even more so now than before.

    Therefore, to represent how well each pitcher performed, the pitcher win should be eliminated. It is misleading. As a simple example, this year John Lackey is 10-12 while Jeremy Hellickson is 12-9. Those numbers tell us neither how well each pitcher performed nor how good a year each had in comparison to the other. In fact, they are perversely misleading.

    True, there are other common stats that fill out the picture, but the upfront one remains the WIN as in 20 GAME WINNER. The issue is how best to replace the win with a simple number that is more representative of reality. I know there is “game score”, but the problem is that calculating it is complex, sort of like quarterback ratings. (Are there game scores for relievers?) When I look at them, I don’t really know what they mean. Perhaps some version of game score could be developed for every pitcher’s performance with clear markers (the way ERA is for many fans, for example) as to what constitutes excellent, good, poor.

    For example, at the end of a season, 3 starters have cumulative game scores of 2560, 1900 and 1250. The scores reflect how many games/innings the pitcher threw plus how well he pitched in those games. Perhaps 2400 (32×75) is considered excellent, 1920 (32×60) is solid and 1440 (32×45) is very poor. We know that A is in contention for the Cy Young, B is a solid mid rotation starter and C had a really poor season.

    Of course there are always caveats. A low total game score may occur because a rookie was called up in June and so pitched fewer games/innings, but did so brilliantly. Or we may have to account for a pitcher’s contribution more subtly if he was used as a long relief/swing man, starting perhaps 8 games and relieving in 30 others. We may want to add average game score to the picture as one corrective.

    And perhaps some similar system can be developed for relievers to eliminate the save and hold while reflecting their performance accurately.

  28. Habeib Khan says:

    Lols Gag is the the Best Lol Network Ever, where you can every thing is lol and Funny, Troll Images, Funny Vidoes, Prank Peoples, Funny Peoples, Prank Images, Fail Pictures, Epic Pictures, Epic Videos, Prank Videos, Fail Videos and Much More Fun and Entertainment, Lols and Gags, Lol Pictures, Lol Videos, Funny Pictures, Lol is the Laugh out of Laugh where you can Fun Unlimited and Laughing Unlimited.

  29. invitro says:

    This revamping of old-fashioned stats is fun. What would be much more fun is if Joe revisits Trout vs. Cabrera when he gets back from tailgating. Because Miggy is down to #5 in WAR in the AL. I’m curious if anyone is still pimping him for MVP. (I’m going with Donaldson.)

  30. Lunk says:

    One aspect of the proposed ERA/RA change that I like is that it addresses something that’s driven me nuts for a long time: pitcher errors.

    I can understand that there’s an argument to be made that CC Sabathia shouldn’t be charged with a run every time Eduardo Nunez confuses Lyle Overbay with an elderly fan sitting three rows behind first. However, there have been a number of times where I’ve seen pitchers have horrific starts that don’t have nearly the effect they ought to on the back of their baseball card because they gave up a bunch of unearned runs after committing an error themselves, as if those runs were somehow not their fault.

  31. JRoth says:

    Is advancing on a wild pitch now an SB? As you say, your goal is to be non-judgmental. A runner on base advances without a hit or walk, it’s clearly a steal.

    Why not?

  32. JRoth says:

    Less nitpicky, and more substantial:

    As I’ve noted before, (and as you kind of acknowledge here), ERA is basically a primitive effort to get at DIPS. So why, in the DIPS era, are we suddenly doing everything possible to make every possible circumstance into a hit*? It would be *more* informative simply to calculate DIO (defense-independent offense) by assigning league-average percentages to LD, GB, and FB. All of these hits and SBs and such that you’re adding may tell a more accurate story (I’d argue that it’s merely more simplistic, not more accurate – the box score tells all as it is), but they don’t aid analysis at all. And what’s the point of advanced stats if they don’t reveal more about the underlying talent?

    *do you explain what happens to dropped 3rd strikes in AVG? You pull a sleight of hand there, saying that it should be a plus for OBP (which I wouldn’t really mind – HBP counts towards OBP as well), but what does it do for AVG? Is it a hit now? Does it go down as a line drive, a ground ball, or a fly ball?

  33. Section 222 says:

    Joe, team wins are pretty easy to count for individual pitchers in Baseball Reference. Just go to the pitcher’s Game Log and look at the results in the games he started. It’s not already calculated or sortable, but at least you can determine the record. And I agree with you it’s alot more useful than the standard win.

  34. I would like for pitcher win stats to at least have the total starts incorporated, like pitcher X is 10-7-9 where 9 is the number of games he started but did not get credit one way or the other. But I’m not sure if that would actually provide any helpful info, except to maybe make a person wonder why a pitcher would have a lot of games he started but did not get a win or loss.

  35. thank’s for your information and i like your post

  36. Habeib Khan says:

    Top 7 Expensive Cars, Top Most Popular Cars, Cars and Vehicles with Info and Pictures, Upcoming Latest cars and vehicles, Latest Mazda Models, Racing Cars, International Sport Cars, Concept Cars, PS-Pod, Strange Vehicles, Nissan, Royce Corniche, Ford Concept Cars, Strange Vehicles, Mercedes and More Sport Cars and Vehicles with Pictures and Info

Leave a Reply

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