My Ideal Metric - TOTAL bases

[BaseballJones]

You need to accumulate four bases to score a run, and obviously those four bases need to be accumulated consecutively. That is, a player who gets four singles with nobody on base in one game doesn't actually accumulate a run for those four bases. But you have to touch first, second, third, and home to score a run. So I'm breaking down runs - which is what ultimately counts in baseball (runs scored and preventing runs from scoring) - into total bases.

The traditional TB metric is simply the number of bases accumulated on hits (one for a single, two for a double, etc.). But I'm including walks, HBP, and base running too. B-ref shows stolen bases and caught stealing (obviously), but also gives us "bases taken" (that is, if you are on first and a single is hit, you get to second automatically, but if you get to third, that's an extra "base taken", for example) and "out on the bases" - so not force plays, but times when you're thrown out, say, trying to score on a possible sac fly. That kind of thing. So that gives us a total bases gained. I also subtract a base when you ground into a double play, since you're removing a runner from the base paths yourself. It's net subtraction. So the formulas are:

TOTAL bases: TB + BB + HBP + SB + BT - (DP + CS + OOB)
TOTAL bases per PA: TOTAL bases / PA

So here are the Sox' leaders in TOTAL bases:

O'Neill: 47
Duran: 40
Casas: 37
Yoshida: 26
McGuire: 21

Leaders in TOTAL bases per PA:

O'Neill: .810
McGuire: .568
Duran: .540
Casas: .522
Hamilton: .522

O'Neil is primarily piling up his bases on homers (7). Duran is a great combo so far of batting TB (26) and net base running TB (+7).

 

[AB in DC]

Interesting. In theory, shouldn't a failed steal of third be negative two bases, since it loses a runner on second? Similarly with outs at third (and three for home).

 

[BaseballJones]

Interesting. In theory, shouldn't a failed steal of third be negative two bases, since it loses a runner on second? Similarly with outs at third (and three for home).

You have a point, but I don't see it that way. Plus, B-ref doesn't tell us what base they were caught stealing at, so I can't easily access that information.

 

[Toe Nash]

You have a point, but I don't see it that way. Plus, B-ref doesn't tell us what base they were caught stealing at, so I can't easily access that information.

https://www.baseball-reference.com/teams/BOS/2024-batting.shtml#all_players_baserunning_batting

 

[Toe Nash]

This sounds similar to "Total Average" but I do like the addition of extra bases taken on hits by others:

https://en.wikipedia.org/wiki/Total_average

 

[BaseballJones]

https://www.baseball-reference.com/teams/BOS/2024-batting.shtml#all_players_baserunning_batting

Oh helpful! Not sure how I missed that. Thanks.

 

[BaseballJones]

Actually, I worked on this formula several years ago and now it hit me why I did DPs and CSs and OOB - not because they relate specifically to Total Bases but because I was also calculating TB per outs made. So a DP is worth two outs, a CS is worth one out, etc. It didn't matter which base you were out on, it just counted as an out regardless.

 

[cantor44]

I've always enjoyed RC/27 - runs created per 27 outs. It largely uses basic conventional numbers, so maybe isn't as esoteric as the current advanced analytical stuff. But it gives you a number for a hitter akin to ERA for a pitcher, and one that makes sense in game terms. I'm surprised it hasn't made it into the mainstream lexicon.

 

[Tuor]

I think this is a great metric. One question though: you mentioned at the start, very appropriately, that the bases have to be consecutive before they really matter. Wouldn’t a HR and the four isolated singles you mentioned count the same in this system after all, though? Is there a way to measure that consecutiveness (consecutivity?)?

 

[DJnVa]

Years ago I worked up a bases/out thing. Similar to this, but CS were -2, etc.

 

[BaseballJones]

I think this is a great metric. One question though: you mentioned at the start, very appropriately, that the bases have to be consecutive before they really matter. Wouldn’t a HR and the four isolated singles you mentioned count the same in this system after all, though? Is there a way to measure that consecutiveness (consecutivity?)?

good question. I’m not really sure how to do that without knowing the actual sequence of events. But it’s kind of like football with yards gained in relationship to points scored by the office. If you get the ball at the 25 yard line, you need to gain 75 yards for a touchdown. The points are what ultimately matters, but the yards are the means by which you get there. The difference between baseball and football of course is that in football you can get the ball anywhere on the field, but in baseball you always begin at home plate.

 

[Tuor]

good question. I’m not really sure how to do that without knowing the actual sequence of events. But it’s kind of like football with yards gained in relationship to points scored by the office. If you get the ball at the 25 yard line, you need to gain 75 yards for a touchdown. The points are what ultimately matters, but the yards are the means by which you get there. The difference between baseball and football of course is that in football you can get the ball anywhere on the field, but in baseball you always begin at home plate.

Yes, that’s a sensible parallel. And of course there are always football offenses that accumulate a lot of yards and score fewer points than it seems they should, and by this metric we might expect a similar phenomenon for some teams to get lots of bases but to strand a lot of runner and score fewer runs than it seems they should.

If anything, that seems useful! Or at least, it would be a number that could help tell a particular story about a team’s performance. I am always leery of “expected” stats — I don’t always trust the assumptions underlying them. But this is a cleaner way to show that same kind of thing. If two teams are about the same in Runs Scored but one has a far higher TOTAL score, that shows something interesting, I think.

 

[luckiestman]

What is the goal of the metric i.e. what question does it intend to answer? I ask because some of these may depend on context vs player control. For example, one team’s philosophy may be to be more aggressive on the base paths and that is not in the individual player’s control.

 

[Njal]

Seems like you would be heading toward wOBA if you start thinking about weighting and consecutivity. wOBA values a homerun plus a strikeout more than two singles.

It's a hard problem because there are pluses and minuses to consecutivity. You generally want your hits bunched together, but not too bunched. If you get all your hits every 10th game, that's bad. But if they're too spread out that's also bad.

EDIT: doing it in a less dumb way, it seems like wOBA likes things to be spread out. Adding additional bases (hr vs triple vs double , etc.) adds at most ~.5 to your wOBA, which is less than what you get by getting on base via an unintentional walk (.69), which is the least valuable way to get on base according to wOBA.

 

[greenmountains]

I think this is a great metric. One question though: you mentioned at the start, very appropriately, that the bases have to be consecutive before they really matter. Wouldn’t a HR and the four isolated singles you mentioned count the same in this system after all, though? Is there a way to measure that consecutiveness (consecutivity?)?

A Home Run is 4 total bases with 1 AB....while 4 singles is 4 total bases in 4 AB's. So a home run is worth 4x a single when accumulated; triple 3x; and double 2x. The question I have is should a base on balls also be considered a AB (for this metric). In isolation, why would a walk (1 base / 0 AB = Infinity) be worth more than a single (1 base / 1 AB = 1.00)?

 

[dirtynine]

This sounds similar to "Total Average" but I do like the addition of extra bases taken on hits by others:

https://en.wikipedia.org/wiki/Total_average

I believe this was the metric that crowned Ken Phelps as an offensive force for a brief time?

 

[BaseballJones]

A Home Run is 4 total bases with 1 AB....while 4 singles is 4 total bases in 4 AB's. So a home run is worth 4x a single when accumulated; triple 3x; and double 2x. The question I have is should a base on balls also be considered a AB (for this metric). In isolation, why would a walk (1 base / 0 AB = Infinity) be worth more than a single (1 base / 1 AB = 1.00)?

I use PA, not AB, so it accounts for walks and HBP.

 

[BaseballJones]

What is the goal of the metric i.e. what question does it intend to answer? I ask because some of these may depend on context vs player control. For example, one team’s philosophy may be to be more aggressive on the base paths and that is not in the individual player’s control.

No metric really takes into account a team's philosophy of running the bases. I mean, if a guy takes off on a called hit and run and the batter doesn't swing, and the runner gets thrown out, is that really his fault? He's only doing what the coach tells him to do, and his teammate failed, but he gets hit with the CS.

So I'm just going by the actual stats themselves.

 

[BaseballJones]

So far in 2024, up til today's action, here were the overall MLB averages. When I use TB, I don't mean batting TB, but adding in walks and HBP and SB and bases taken on the base paths.

TB / PA = 0.494
TB / OUT = 0.693
TB / R = 4.10

Here's Boston's #s:

TB / PA = 0.490
TB / OUT = 0.679
TB / R = 4.14

So the Sox are below average in all three metrics.

Here's where the Sox' players are with these:

 

[Tuor]

A Home Run is 4 total bases with 1 AB....while 4 singles is 4 total bases in 4 AB's. So a home run is worth 4x a single when accumulated; triple 3x; and double 2x. The question I have is should a base on balls also be considered a AB (for this metric). In isolation, why would a walk (1 base / 0 AB = Infinity) be worth more than a single (1 base / 1 AB = 1.00)?

Yes, and (acknowledging Jones's clarification about PA v AB) the TOTAL/PA version of the stat would account for that difference between HR and singles. But my point was more focused on the counting version. The guy who goes 1 for 4 with a solo HR and the guy who goes 4 for 4 with 4 singles and is stranded every time both end the game with the same TOTAL score. One has created a run, of course, and the other has not. It seems a weakness in the stat in its ability to fulfill the initially stated goal of "breaking down runs into total bases." But I do accept the football parallel mentioned before. Getting bases and scoring runs are not, in fact, exactly the same, any more than accumulating offensive yards and scoring points are the same in football. But, despite some interesting anomalies, they are doubtless likely to be highly correlated, so I do like this stat.

Jones, I hope you continue to update this thread with your calculations. It will be interesting to see how, over the course of a live season, this metric weighs the offensive contributions of Sox players as compared to other metrics.

 

[Tuor]

It needs a better abbreviation, though! TotB? "TrueTB" is kind of fun, but I think perhaps a different word than "total" might be better to differentiate it from TB.

 

[zenax]

The question is with a new 'stat' is how well it predicts something.

 

[Tuor]

The question is with a new 'stat' is how well it predicts something.

Personally, I disagree. I obviously understand why GM's want that, but that isn't what I look for in a stat. For me, the question with a new stat is the extent to which it helps us to perceive a more nuanced narrative. To use a simple example, the shift from the old days when BA was the main thing to the days when people started caring about OPS was interesting and useful to me because it enabled a more nuanced understanding of a player's offensive performance -- OPS told more of the story than BA by itself, and it told it better, especially since you could look at both numbers at once. You can intuit what a "high BA + low OPS" means, what it translates to in the narrative of a player's performance. I like this new stat because it embraces all elements of offense -- batting plus base-running -- in a simple and intuitive way. That's lovely, and really useful -- not just for potentially projecting the future, but for appreciating and understanding the story that I am currently watching unfold: the 2024 Red Sox season. Many fans enjoy the focus on the future and approach the game like amateur GM's. That's great, and I hope they get enjoyment out of that, but I am in it for the present, for the enjoyment of the stories that we are watching unfold. And this stat will help me, I think, to notice things I would not have noticed and to appreciate things I might not have observed.

For instance, see how in Jones's table above Casas and Duran are very close in TOTAL/PA and TOTAL/OUT? Casas and Duran are very different offensive players, and it is hard to find a clear and objective way to compare their real offensive impact on the team. This stat shows how Casas's "walks plus majestic bombs" input has had almost exactly the same impact on the Sox's offense as Duran's "singles + havoc" input. And I love that! It would be very hard to compare them clearly otherwise.

 

[RS2004foreever]

You need to accumulate four bases to score a run, and obviously those four bases need to be accumulated consecutively. That is, a player who gets four singles with nobody on base in one game doesn't actually accumulate a run for those four bases. But you have to touch first, second, third, and home to score a run. So I'm breaking down runs - which is what ultimately counts in baseball (runs scored and preventing runs from scoring) - into total bases.

The traditional TB metric is simply the number of bases accumulated on hits (one for a single, two for a double, etc.). But I'm including walks, HBP, and base running too. B-ref shows stolen bases and caught stealing (obviously), but also gives us "bases taken" (that is, if you are on first and a single is hit, you get to second automatically, but if you get to third, that's an extra "base taken", for example) and "out on the bases" - so not force plays, but times when you're thrown out, say, trying to score on a possible sac fly. That kind of thing. So that gives us a total bases gained. I also subtract a base when you ground into a double play, since you're removing a runner from the base paths yourself. It's net subtraction. So the formulas are:

TOTAL bases: TB + BB + HBP + SB + BT - (DP + CS + OOB)
TOTAL bases per PA: TOTAL bases / PA

So here are the Sox' leaders in TOTAL bases:

O'Neill: 47
Duran: 40
Casas: 37
Yoshida: 26
McGuire: 21

Leaders in TOTAL bases per PA:

O'Neill: .810
McGuire: .568
Duran: .540
Casas: .522
Hamilton: .522

O'Neil is primarily piling up his bases on homers (7). Duran is a great combo so far of batting TB (26) and net base running TB (+7).

It's been a while, but isn't runs created pretty close to that? You could argue DP's are a product of factors beyond the hitters control and is a little like RBI's.
I went back and ran a number very close to what you are suggesting.

 

[BaseballJones]

It's been a while, but isn't runs created pretty close to that? You could argue DP's are a product of factors beyond the hitters control and is a little like RBI's.
I went back and ran a number very close to what you are suggesting.

I don’t know how runs created is calculated but the reason I’m sticking with bases is because runs created gives the impression that those are ACTUAL runs created and not just theoretical runs created. But they’re not. I want to stick to the constituent parts of run-making and not translate it into hypothetical runs created.

 

[mauf]

The question is with a new 'stat' is how well it predicts something.

Forward-looking and backward-looking stats each have their purpose. The analytics movement has traditionally focused on forward-looking stats because (1) the people who have a financial interest in getting smarter (front offices, gamblers, et al.) are more interested in predicting what’s next than in analyzing what already happened; and (2) rightly or not, there has been a general sense that Bill James-era advanced stats (runs created, et al.) did the backward-looking thing well enough. But that doesn’t mean a new stat that is backward-looking can’t add value, and we shouldn’t dismiss it simply because it isn’t valuable as a predictive tool.

 

[pjheff]

It needs a better abbreviation, though! TotB? "TrueTB" is kind of fun, but I think perhaps a different word than "total" might be better to differentiate it from TB.

A JimEd? A Rice?

 

[zenax]

Forward-looking and backward-looking stats each have their purpose. The analytics movement has traditionally focused on forward-looking stats because (1) the people who have a financial interest in getting smarter (front offices, gamblers, et al.) are more interested in predicting what’s next than in analyzing what already happened; and (2) rightly or not, there has been a general sense that Bill James-era advanced stats (runs created, et al.) did the backward-looking thing well enough. But that doesn’t mean a new stat that is backward-looking can’t add value, and we shouldn’t dismiss it simply because it isn’t valuable as a predictive tool.

Doesn't it need to be tied to league stats to be useful? Go back to a year like 1930. The NL: .303/.360/.448/.808 and AL: .288/.351//.421/.772, yet the A's beat the Cardinals 4 games to 2 in the World Series, outscoring them 21 to 12.

i know, my my argument above...stuff happens, small sample size, but when you talk about stat measures, don't they need to be backwardly predictive? How can you compare players from different eras, otherwise?

 

[mauf]

Doesn't it need to be tied to league stats to be useful? Go back to a year like 1930. The NL: .303/.360/.448/.808 and AL: .288/.351//.421/.772, yet the A's beat the Cardinals 4 games to 2 in the World Series, outscoring them 21 to 12.

i know, my my argument above...stuff happens, small sample size, but when you talk about stat measures, don't they need to be backwardly predictive? How can you compare players from different eras, otherwise?

Well, a statistic needs to measure something, and the jury’s out as to whether the OP’s proposed metric does that. I’m just saying it doesn’t necessarily need to be predictive; there’s value in accurately quantifying what already happened.

 

[Hank Scorpio]

I wonder if someone could come up with a sort of “weighted” total bases. Because I think a leadoff triple isn’t exactly three times as valuable as a walk or a single. A runner on third can score on a flyball, or a single. A runner on second can advance to third on a ground out. But a runner on first can be erased easily with a double play, and is not going to score on one or two outs - or a base hit, the way a runner on second and third are.

Figuring it might be a bit… involved.

 

[BaseballJones]

Well, a statistic needs to measure something, and the jury’s out as to whether the OP’s proposed metric does that. I’m just saying it doesn’t necessarily need to be predictive; there’s value in accurately quantifying what already happened.

Well the statistic clearly measures something. That's not in question. It's whether that something is worth measuring. I'm very much open to criticism on that point. And I agree that legit metrics don't have to be predictive, they can just measure what did happen. I'm not sure that this couldn't be predictive in some way though.

 

[Petagine in a Bottle]

I wonder if someone could come up with a sort of “weighted” total bases. Because I think a leadoff triple isn’t exactly three times as valuable as a walk or a single. A runner on third can score on a flyball, or a single. A runner on second can advance to third on a ground out. But a runner on first can be erased easily with a double play, and is not going to score on one or two outs - or a base hit, the way a runner on second and third are.

Figuring it might be a bit… involved.

Isn't that this?

https://library.fangraphs.com/principles/linear-weights/

 

[Hank Scorpio]

Isn't that this?

https://library.fangraphs.com/principles/linear-weights/

Well, now I know what the w in wOBA means.

Meanwhile in another thread…

Hank: “someone should invent like an edible food holder… like you can put food in between it, and just eat a bunch of stuff at once that would otherwise be messy… I’m not sure what it should taste like though.”

“Bread Hank. It’s called bread. It exists. And it’s called bread.”

 

[Sin Duda]

I was really encouraged to see the MassLive article on Houck's performance in his "Maddox" to say this: "Houck has made an effort to not overthrow this season. With all the focus in the game on velocity, that’s been hard at times. But he’s found that the less he exerts himself on the mound, the more in control he feels. The adjustment has resulted in somewhat lower velocity on his fastball, but it’s also enabled him to stay within the strike zone more regularly."

 

[zenax]

Has anyone looked at Team Win Probability (bb-ref)? It is available by players with a team under Detailed Stats when you select Batting (Pitching)/Detailed Stats/Team Win Probabilty.

For example: Team Win Probability (batting -- Red Sox 2024)