Bad Stats

[element1286]

I will agree that wins is a little incomplete... there is 2 other stats in the Wins category that would help the win/loss stat

1. Win/loss record in close games(and percentage)
2. win/loss record of team in games pitched(and percentage)

With those 2 extra win stats you can make a better understanding to the value of the pitchers wins...

As for saves i think save percentage is a very important stat... isnt that the whole purpose of a closer??

What are you measuring with "wins" that you don't measure with how many runs a pitcher gives up (ERA) or how many runs a pitcher is responsible for (SIERA, FIP)?

 

[Voltaire26]

I will agree that wins is a little incomplete... there is 2 other stats in the Wins category that would help the win/loss stat

1. Win/loss record in close games(and percentage)
2. win/loss record of team in games pitched(and percentage)

With those 2 extra win stats you can make a better understanding to the value of the pitchers wins...

As for saves i think save percentage is a very important stat... isnt that the whole purpose of a closer??

Saves can be gotten for retiring one batter or 3 outs if you team is ahead by three. Kinda cheap in my opinion. However, that should not diminish Mariano Rivera's career and he certainly be a 1st year Hall of Famer (if not unanimous).

 

[MilkSpiller22]

What are you measuring with "wins" that you don't measure with how many runs a pitcher gives up (ERA) or how many runs a pitcher is responsible for (SIERA, FIP)?

dont understand your question... a win is measured by if a pitcher comes out of the game with the lead(if pitcher starts the game then after the 5th inning minimum) and the team wins the game without ever losing the lead...

I know you knew that, but i dont know how to answer your question...

 

[navamind]

I don't think there's nec anything wrong with most stats, just so long as you know what they mean - or don't mean. And some stats are better at predicting things while other stats are better at describing what had happened. I generally think fans overrate rate-stats while ignoring IP or PA and defensive stats are nearly useless. I think they can tell us correctly who is good/bad but putting an accurate value on that person is nearly impossible.

Absolutely. Sample size matters.

 

[element1286]

dont understand your question... a win is measured by if a pitcher comes out of the game with the lead(if pitcher starts the game then after the 5th inning minimum) and the team wins the game without ever losing the lead...

I know you knew that, but i dont know how to answer your question...

My question is why measure wins and not just runs a pitcher is responsible for or given up. Unless there is an underlying hypothesis that giving up 4 runs in situation A is equal to giving up 3 runs in situation B, you don't need to go through the machinations of situational stats.

 

[Lemon Harang Pie]

Can you explain this reasoning, I understand having reservations with WAR because of defense and replacement value, but I can't think of any reason OPS+ is superior to wRC+.

I just don't see any real advantage to catch-all statistics designed to essentially rank the entirety of what is done on the field and see some real drawbacks.

OPS+ is often used incorrectly but when used properly I think it gives a decent idea of where two different hitters stand. The biggest drawback when used properly is I just don't know how accurate Park Factors really is.

 

[Lemon Harang Pie]

Absolutely. Sample size matters.

Penis size too. Anyone who says otherwise is a little guy.

 

[MilkSpiller22]

My question is why measure wins and not just runs a pitcher is responsible for or given up. Unless there is an underlying hypothesis that giving up 4 runs in situation A is equal to giving up 3 runs in situation B, you don't need to go through the machinations of situational stats.

what i am saying though, is that we measure wins... but we have done it in such a way that we have no clue if it actually says anything or not... by creating these 2 subsets of the win category we can actually explore the importance of the category, whether it truly is entirely a "team" stat or if there actually is truth to some pitchers find ways to win while others find ways to lose...

 

[element1286]

I just don't see any real advantage to catch-all statistics designed to essentially rank the entirety of what is done on the field and see some real drawbacks.

OPS+ is often used incorrectly but when used properly I think it gives a decent idea of where two different hitters stand. The biggest drawback when used properly is I just don't know how accurate Park Factors really is.

I'm not sure I understand, how is OPS+ not a catch-all statistic? I understand looking at OBP and SLG separately, but that isn't OPS, it's just OBP and SLG.

 

[element1286]

what i am saying though, is that we measure wins... but we have done it in such a way that we have no clue if it actually says anything or not... by creating these 2 subsets of the win category we can actually explore the importance of the category, whether it truly is entirely a "team" stat or if there actually is truth to some pitchers find ways to win while others find ways to lose...

If you want to measure that I understand, but I don't see how your proposal does so. When you measure anything involving a team accomplishment, ie win, and then reducing it down to a specific players, how are you separating the extraneous variables like team offense, team bullpen, team defense?

 

[MilkSpiller22]

If you want to measure that I understand, but I don't see how your proposal does so. When you measure anything involving a team accomplishment, ie win, and then reducing it down to a specific players, how are you separating the extraneous variables like team offense, team bullpen, team defense?

Still keeping it simple, keeping statistics simple is important even when you are complicating them...

Dont want to measure things that are too hard to measure... especially when it is irrelevant...

 

[element1286]

Still keeping it simple, keeping statistics simple is important even when you are complicating them...

Dont want to measure things that are too hard to measure... especially when it is irrelevant...

I agree, but some problems can't be solved with simplicity. I think this is one of those situations.

 

[MilkSpiller22]

I agree, but some problems can't be solved with simplicity. I think this is one of those situations.

If you take away unwinnable games(when your team scores 0 runs for you) then wont the rest of the situationals even out as the sample size gets larger??

 

[Lemon Harang Pie]

I'm not sure I understand, how is OPS+ not a catch-all statistic? I understand looking at OBP and SLG separately, but that isn't OPS, it's just OBP and SLG.

It might be a matter of perspective but I think the two are used very differently.

OPS+ is just kind of an easier way to look at OBP and SLG in relation to the rest of the league whereas wRC+ puts definitive values on hits, walks, etc. to come up with a more definitive answer. Definitive answers are fine in vacuums but the games aren't played in vacuums.

 

[MilkSpiller22]

It might be a matter of perspective but I think the two are used very differently.

OPS+ is just kind of an easier way to look at OBP and SLG in relation to the rest of the league whereas wRC+ puts definitive values on hits, walks, etc. to come up with a more definitive answer. Definitive answers are fine in vacuums but the games aren't played in vacuums.

OPS+ is misused all the time... people do use it as the tell all stat... and that is not the way to use it... as it is just a adjusted OPS... But i am not a fan of it since i dont like adjusted stats... atleast not how they adjusted it... Never a they fan when it comes to stats...

 

[$500 Million]

I don't think there are any "bad stats", there are just stats that some like better than others. It's a good thing there are a ton of them, we can all pick and choose our own.

 

[gunnarthor]

Penis size too. Anyone who says otherwise is a little guy.

Or has passed his oral exams.

 

[element1286]

It might be a matter of perspective but I think the two are used very differently.

OPS+ is just kind of an easier way to look at OBP and SLG in relation to the rest of the league whereas wRC+ puts definitive values on hits, walks, etc. to come up with a more definitive answer. Definitive answers are fine in vacuums but the games aren't played in vacuums.

Eh, I still don't get your view, why would you want less precision? Why not take the most precise measure we have and use the implicit assumption that we don't know everything about how to win baseball games.

 

[element1286]

If you take away unwinnable games(when your team scores 0 runs for you) then wont the rest of the situationals even out as the sample size gets larger??

No, because results that are team based are not normally distrusted over the population. You'd have to measure guys with similar team environments or guys that have changed teams recently.

 

[Lemon Harang Pie]

Eh, I still don't get your view, why would you want less precision? Why not take the most precise measure we have and use the implicit assumption that we don't know everything about how to win baseball games.

Precision isn't that important without accuracy.