For Stats Guys

[MARCPELLETIER]

your formula will overvalue power guys who walk a lot, but will undermine the Tony-Gwynn/Wade-Boggs type.

Basically, in your formula, a single = a walk (bad).

Also 1 HR+ 3 outs = 4 singles (really bad)

I would at least add hits in the left part of your formula, and then multiply BB by 1.5 (only in the left side of your formula)

hits+TB+1.5*BB+1.5*HBP+...

That being said, a very, very good formula, and simple, is to calculate:

1.2*onbase + slugging.

[The Turtle]

hey Marcus, why the 1.2? (i know this may be a complex answer, but hopefully not)

[quote:9066951f80]
I would at least add hits in the left part of your formula, and then multiply BB by 1.5 (only in the left side of your formula)

hits+TB+1.5*BB+1.5*HBP+...[/quote:9066951f80]

this is getting more complex and I am trying to find a formula easy like OPs but stronger

And i don't know if if it undervalues Boggs and Gwynn types ..i got looking at this due to the 1987 MVP debate which was Boggs best season and he rates better than Trammell and G. Bell in that season (although he did hit 24 HRs , but thats vs 49 for Bell) If i do understate hitters like Ichiro using Base Average then Boggs was Far and Away better

[Mean Dean]

Something extremely similar to this stat exists: [url=http://www.baseball-statistics.com/Leaders/glossary.htm]Total Average[/url]. It's actually on your league leaderboard here, and is calculated for every player by the CD-ROM. There's no point to it, though. You still need a calculator to figure it out, and it's less accurate than OPS, RC, etc. If you're making a stat, you can't just add up a bunch of numbers in a way that "looks right." Sure, it might "make sense to people", but if it's not accurately describing what offense is, who cares? What you have to do is see how your formula correlates to real-life team run scoring, and compare that to other formulae. That's the test.

[The Turtle]

[quote:c632ee08fc]accurately describing what offense is[/quote:c632ee08fc]

that simply has not been able to be done....nor am I looking for it, what I am looking for is an easy method that is better than OPS and I can't see why OPS is a better indiactor than seeing how many bases a player hits for per AB.

Is it the fact that it adds hits twice?? per Luckyman's post I might tend to think he would lean that way

[Mean Dean]

[quote:3527715ed7]that simply has not been able to be done[/quote:3527715ed7]What do you mean? Runs created predicts team run scoring within a margin of error of around 20 runs.

Why do runs created and similar stats work better than Total Average and similar stats (and for that matter, OPS)? I think it boils down to the fact that they multiply the elements of offense rather than adding them. As James put it:[quote:3527715ed7]The creation of runs is not a linear activity, in which each element of the offense has a given weight regardless of the situation, but rather a geometric activity, in which the value of each element is dependent on the other elements. Linear weighting treats the elements of an offense as if they were discrete units, each bearing no relationship to the other; runs created varies the value of each according to its relationship to the others... By runs created, the incremental value of a home run could be as high as 1.60 runs (if the team has a lot of people on base and comparatively little power) or as low as 1.30 (if the team has very few people on base but already has a lot of power.)[/quote:3527715ed7]But honestly, it doesn't matter why it works better unless you're a sabermetrician. The point is that it does work better. If you know how to do a correlation test, you can check it for yourself. Total Average will correlate worse with real-life team run scoring than OPS, which in turn will correlate worse than RC.

[The Turtle]

TOTAL AVERAGE Tom Boswell's formulation for offensive contribution from a variety of batting and baserunning events. The concept of the numerator is bases gained, that of the denominator is outs made:

(Total Bases + Steals + Walks + HBP - Caught Stealing)
-----------------------------------------------------
(At-Bats - Hits + Caught Stealing + GIDP)


very similar to what i was thinking

[The Turtle]

[quote:363936e0b9]f you know how to do a correlation test[/quote:363936e0b9]

exactly the opposite of what my post is intending to do, i'm looking for simple here not a complex formula that accurtely depicts runs using linear weights i know all this stuff, but I am looking for something even easier and I am trying to get some responses outside of what Tango Tiger and James have told us

something more accurate than OPS, perhaps not as accurate as RC, and still easy enough for Joe Blow (I think RBI's are an important stat) kinda person

[The Turtle]

[quote:7c886aa63c]Total Average will correlate worse with real-life team run scoring than OPS[/quote:7c886aa63c]

perhaps a reason for this would be nice, any one have a correlation test lying around so I don't have to do it ??? :)

I'm not trying to push this formula as valid mind you just trying to expound a reason why its not and then try to fix something easier than linear weights for people who have trouble even understanding the importance of OBP (like my friend who works for thew Astros)

[Mean Dean]

[quote:0ed7f3ff66]perhaps a reason for this would be nice[/quote:0ed7f3ff66]That quote explains why. Addition isn't the best way to simulate how an offense works; you need that multiplication element, to reflect that the elements of offense interact with each other.[quote:0ed7f3ff66]any one have a correlation test lying around so I don't have to do it ??? [/quote:0ed7f3ff66]In Excel, if you for instance had the team runs scored in cells A1 through A30 and the team runs created in cells B1 through B30, you would do:[code:1:0ed7f3ff66]=CORREL(A1:A30,B1:B30)[/code:1:0ed7f3ff66]

[MARCPELLETIER]

1.2 is there only to provide a bit more importance to on-base vs slugging in the ops formulation.