Odd SP out

[drivlikejehu]

 

Mind your manners.  There is nothing about this question that is going to be proven "factually".  It will be each person's opinion which is just fine, including yours.

 

May is better at getting hitters out. He hasn't been as lucky in ERA terms as Pelfrey or Milone. It is what it is. I don't see much reason to pretend that a factually incorrect understanding of performance should be considered a legitimate opinion.

[TheLeviathan]

 

May is better at getting hitters out. He hasn't been as lucky in ERA terms as Pelfrey or Milone. It is what it is. I don't see much reason to pretend that a factually incorrect understanding of performance should be considered a legitimate opinion.

 

I don't think we need to defend May's worthiness with such hostility.  Some people value past actual results over indicators of future results.  It doesn't make their opinion illegitimate.

[drivlikejehu]

 

I don't think we need to defend May's worthiness with such hostility.  Some people value past actual results over indicators of future results.  It doesn't make their opinion illegitimate.

 

The discussion is over who should stay in the rotation, which relates only to future performance. If you want to value past luck, that's one thing, but it is absolutely a fact that Pelfrey and Milone's luck is not predictive.

[nicksaviking]

 

I do think it's close, but Milone's numbers are the best. In a vacuum, I'd take the Pelfrey line before May's.

Milone 3.19 ERA 1.29 WHIP

May     4.37 ERA 1.37 WHIP

Pelfrey  3.81 ERA 1.44 WHIP

 

 

 

I saw May in a jam with runners on the corners and no outs three starts ago and bases loaded and no outs last start.  In those two innings combined he didn't give up a run simply because he got five of the six needed outs on strikeouts.

 

Milone and Pelfrey can't do that. 

 

Also, May from this spring is not the same pitcher as May since June.  He has been pretty damn good lately.  Which of course is another reason not to knock him out of the pen, he still has room to improve, Milone and Pelfrey are already at their ceilings.

[USAFChief]

The discussion is over who should stay in the rotation, which relates only to future performance. If you want to value past luck, that's one thing, but it is absolutely a fact that Pelfrey and Milone's luck is not predictive.

It's neither factual that the past performance of each is "luck," nor that the past isn't predictive.

Those are called assertions, or theories, not facts.

 

Luck MIGHT be a factor. I'd buy PROBABLY is a factor. IS the only factor? No. This is baseball. Neither you, I, or the folks at Fangraphs have completely figured out why things happen.

 

And we sure as heck can't say for a fact what is, and isn't going to happen in the future. Past performance is what everyone uses tp predict the future...including you. You're just assigning different values to different parts of past performance.

 

I think May should stay in the rotation, probably ahead of Pelfrey. But it's absolutely a close call at this point, and there are reasonable arguments for any of the three.

 

That's the "facts," as I see them.

[Craig Arko]

We'll know soon enough. Santana has to be activated tomorrow even though he won't be pitching until Sunday.

[drivlikejehu]

 

It's neither factual that the past performance of each is "luck," nor that the past isn't predictive.
Those are called assertions, or theories, not facts.

Luck MIGHT be a factor. I'd buy PROBABLY is a factor. IS the only factor? No. This is baseball. Neither you, I, or the folks at Fangraphs have completely figured out why things happen.

And we sure as heck can't say for a fact what is, and isn't going to happen in the future. Past performance is what everyone uses tp predict the future...including you. You're just assigning different values to different parts of past performance.

I think May should stay in the rotation, probably ahead of Pelfrey. But it's absolutely a close call at this point, and there are reasonable arguments for any of the three.

That's the "facts," as I see them.

 

The predictive value of various statistical measures has been evaluated empirically. The lack of predictive value for short-term ERA is absolutely a fact. No case against this fact has ever been presented. If you can do so, or point to where someone else has, then by all means do so. Otherwise your objections are per se invalid.

 

I was also only commenting on one aspect of the May v. Pelfrey v. Milone comparison. Even if Pelfrey or Milone did have a performance argument, it would be completely outweighed by May's long-term value, which is best served by keeping him in the rotation.

 

No rational case can be made to remove May from the rotation. Some questions in baseball are subjective and impossible to resolve definitely. This is not such a question. Any conclusion other than May staying in the rotation is erroneous, period.

[TheLeviathan]

 

The discussion is over who should stay in the rotation, which relates only to future performance. If you want to value past luck, that's one thing, but it is absolutely a fact that Pelfrey and Milone's luck is not predictive.

 

ERA is 100% unable to predict future performance?  That's new - you got a link for that?

 

Because if it's not 100% unable to predict it, than one could value the stat.  There might be other stats that have proven, more often than not, to be better predictors, but none of them are certainties.

 

Quit talking like they are.  This kind of tone with metrics is why so many of them have trouble catching on.  Seriously, back down a bit.

[Brock Beauchamp]

May is better at getting hitters out. He hasn't been as lucky in ERA terms as Pelfrey or Milone. It is what it is. I don't see much reason to pretend that a factually incorrect understanding of performance should be considered a legitimate opinion.

It is a legitimate opinion. There's no need to be hostile with someone over preferred methods of analyzing player performance. It's not like he's countering your advanced metrics with "wins and grit".

[USAFChief]

 

Even if Pelfrey or Milone did have a performance argument, it would be completely outweighed by May's long-term value, which is best served by keeping him in the rotation.

 

  Any conclusion other than May staying in the rotation is erroneous, period.

Two more examples of "things that are not facts."

 

These are best termed "opinions."

[drivlikejehu]

 

ERA is 100% unable to predict future performance?  That's new - you got a link for that?

 

Because if it's not 100% unable to predict it, than one could value the stat.  There might be other stats that have proven, more often than not, to be better predictors, but none of them are certainties.

 

Quit talking like they are.  This kind of tone with metrics is why so many of them have trouble catching on.  Seriously, back down a bit.

 

A number of knowledgeable fans such as yourself go off course when it comes to some of the finer statistical points regarding ERA and it's defense-independent compatriots. The predictive value of ERA is all about sample size - at a certain point, it becomes the best predictor of future ERA.

 

In a sample sample, however, it has zero value. Not because the small-sample ERA tells us literally nothing - for instance, even Milone's ERA in 53 IP is enough to be confident he isn't actually a mediocre high schooler that is just getting lucky.

 

But, all of the valuable information being captured by ERA is also captured by xFIP. I've never used an underline before on TD, but that is really the crucial point. You are never better off using small sample ERA as opposed to xFIP, because the expected error value is always higher with ERA than with xFIP.

 

Therefore, a rational analyst must use a normalized metric over ERA in a small-sample situation to produce the lowest expected error. I am at a loss as to how any other conclusion can be reached. If you object to my tone, I understand, but as to my conclusion I would appreciate a substantive explanation of how any other outcome is possible.

[TheLeviathan]

 

Therefore, a rational analyst must use a normalized metric over ERA in a small-sample situation to produce the lowest expected error. I am at a loss as to how any other conclusion can be reached. If you object to my tone, I understand, but as to my conclusion I would appreciate a substantive explanation of how any other outcome is possible.

 

Your tone is the problem.  I have no problem with the conclusion "xFIP is better for predicting".  It is better.  But it is not "irrational" or "erroneous" or "objectively wrong" or "invalid" to use ERA.

 

Using ERA over xFIP is just very likely to lead to more errors, not certain to always lead to errors.  That's where you're going wrong and it's why your tone is coming off so ridiculously.

[drivlikejehu]

 

Your tone is the problem.  I have no problem with the conclusion "xFIP is better for predicting".  It is better.  But it is not "irrational" or "erroneous" or "objectively wrong" or "invalid" to use ERA.

 

Using ERA over xFIP is just very likely to lead to more errors, not certain to always lead to errors.  That's where you're going wrong and it's why your tone is coming off so ridiculously.

 

Flipping a coin may also lead to correct outcomes. My point was about the decision-making process. Decisions can only be made on the basis of available information. 

 

Since, as you admit, the expected result will be better when using xFIP, as opposed to ERA, it would be wrong to use ERA. You would intentionally be using a method that, on the whole, produces inferior results.

 

So can you explain why it is valid to use ERA, knowing that it is inferior?

[TheLeviathan]

 

it would be wrong to use ERA. You would intentionally be using a method that, on the whole, produces inferior results.

 

So can you explain why it is valid to use ERA, knowing that it is inferior?

 

Just because something isn't the most prudent method doesn't make it invalid.  You're speaking of likelihoods in the language of certainties.

[drivlikejehu]

 

Just because something isn't the most prudent method doesn't make it invalid.  You're speaking of likelihoods in the language of certainties.

 

The comments of mine you object to are all based on a crucial distinction you aren't recognizing.

 

Decisions are based on expected future results, not actual future results. With respect to expected future results, one should always use xFIP in the situation under discussion. 

 

Using the correct approach does not always pay off. In hold 'em poker, being dealt 2 aces justifies going all-in without hesitation, but a loss can still easily occur. Does that make it valid to fold those 2 aces when raised?

 

It is a certainty that you should give yourself the best chance to succeed. That was my point. You have agreed with all the underlying facts, except for holding out that using an inferior method is still justifiable. That's not a position that can be reconciled with good decision making.

[TheLeviathan]

 

Using the correct approach does not always pay off. In hold 'em poker, being dealt 2 aces justifies going all-in without hesitation, but a loss can still easily occur. Does that make it valid to fold those 2 aces when raised?

 

And you seem to be missing there is a wide range of options between "all-in" and "fold".

[drivlikejehu]

 

And you seem to be missing there is a wide range of options between "all-in" and "fold".

 

Not in the scenario I was implying.

 

Heads up, equal stacks. You are big blind and dealt 2 aces. Other player raises all-in. Is it justifiable to fold?

[TheLeviathan]

 

Not in the scenario I was implying.

 

Heads up, equal stacks. You are big blind and dealt 2 aces. Other player raises all-in. Is it justifiable to fold?

 

Nope, but I might slow play.  Again, you're talking probabilities in the language of certainties.  Your method may be more reliable, that doesn't make it "irrational" to try something with less probability but still some success.

[drivlikejehu]

 

Nope, but I might slow play.  Again, you're talking probabilities in the language of certainties.  Your method may be more reliable, that doesn't make it "irrational" to try something with less probability but still some success.

 

You can't slow play when your opponent goes all-in pre-flop.

 

I think your claim here, which I find very surprising, is that it is rational to knowingly make poor decisions. To me, that seems like the opposite of rationality, which is just "the exercise of reason." I suppose that is the core of the disagreement. I think it is irrational to intentionally make poor decisions, in the hope of getting lucky.

[Linus]

Or I might be the pitching coach who, working with all involved everyday, thinks pitcher A is throwing better than pitcher B right now. Or, stats no matter how sophisticated are not the only decision making factors.

[jorgenswest]

Does anyone have an opinion on whether the Twins should start Gibson Tuesday and give him two more starts? Doing so would take a start away from May, Pelfrey or Milone.

[USAFChief]

 

Not in the scenario I was implying.

 

Heads up, equal stacks. You are big blind and dealt 2 aces. Other player raises all-in. Is it justifiable to fold?

We're not preflop any more.  We're halfway through the season...in poker terms, we've at least seen the flop.

 

At that point, your two aces may very well be behind, and calling an all in might be your last hand.

 

Besides...the Twins have TWO aces??  

[TheLeviathan]

 

Does anyone have an opinion on whether the Twins should start Gibson Tuesday and give him two more starts? Doing so would take a start away from May, Pelfrey or Milone.

 

In my opinion, Pelfrey should lose his spot.  He's had a great couple months, but he's the right choice.

[drivlikejehu]

 

We're not preflop any more.  We're halfway through the season...in poker terms, we've at least seen the flop.

 

 

I only used the analogy with respect to whether there is such thing as a right or wrong decision. Multiple posters are claiming that it's OK to use incorrect reasoning because maybe things will turn out OK anyway. 

[TheLeviathan]

 

I only used the analogy with respect to whether there is such thing as a right or wrong decision. Multiple posters are claiming that it's OK to use incorrect reasoning because maybe things will turn out OK anyway. 

 

You are claiming it to be "irrational", I don't think you understand what that word means.  Using ERA is unwise, not irrational.  But I don't think anyone can explain this to you, you've more than demonstrated that.

[USAFChief]

 

I only used the analogy with respect to whether there is such thing as a right or wrong decision. Multiple posters are claiming that it's OK to use incorrect reasoning because maybe things will turn out OK anyway. 

I think multiple posters are claiming XFIP might not be the only piece of information that properly goes into making a decision like this.

 

 

[Linus]

I think multiple posters are claiming XFIP might not be the only piece of information that properly goes into making a decision like this.

This. The only "irrational" part of this discussion is to be so wedded to any particular statistic that you would completely disregard any other factor. And then to do it with a condescending attitude makes for a snarky thread.

[drivlikejehu]

 

You are claiming it to be "irrational", I don't think you understand what that word means.  Using ERA is unwise, not irrational.  But I don't think anyone can explain this to you, you've more than demonstrated that.

 

I know what it means. But here's what dictionary.com says:

 

1. without the faculty of reason; deprived of reason.
2. without or deprived of normal mental clarity or sound judgment.
3. not in accordance with reason; utterly illogical

... 

 

The first and second generally would apply to individuals, whereas I was referring to an abstract idea, which fits well with the third definition above.

 

Relying on small sample size ERA is not in accordance with reason and is therefore irrational.

 

Before you question the use of "reason," I used it in the context that dictionary.com provides as definition #4: "sound judgment; good sense."

 

Relying on small sample size ERA is not in accordance with sound judgment and is therefore irrational.

[Brock Beauchamp]

I only used the analogy with respect to whether there is such thing as a right or wrong decision. Multiple posters are claiming that it's OK to use incorrect reasoning because maybe things will turn out OK anyway.

The problem with metrics (I generally agree with your thinking on this subject) is that they cannot equate for the human factor of baseball. Sometimes, that counts and is displayed in the box scores of games. Metrics are good predictors of future performance but they're not the only reason a decision should be made. Sometimes, a guy has a hot hand and you just go with it, metrics be damned.

 

In the long game, the metrics will prove more accurate. Right now, the Twins aren't playing a long game.

 

I'm not saying that's the case in this situation because I have no evidence to support it... People who are on-hand have to make that kind decision. I only hope they make the right call.

[drjim]

After all this, I still think May stays and Pelfrey goes to the bullpen and we can all be happy.