A win clinches you a spot in the playoffs...

[WATERY]

It's interesting here, because what your actually arguing for HGM is a return to the way reliever use (basically) used to be. It used to be that teams didn't have a "closer", they actually had a "Fireman". Basically, each team selected (usually one) ace reliever, who would come into the game to "put out fires"; hence the term "fireman". Very retro viewpoint, really.


I'm actually surprised that you didn't receive more supportive posts on this one. Starting out making it more personally about Torre and the Yanks probably didn't help much, but it's still somewhat surprising. Returning to the "fireman" style of reliever use is a much more traditionalist viewpoint... although, that seems to be changing quickly. There's a chapter in Baseball Behind the Numbers which, while not actually advocating it, is supportive of it. I hear that Bill James and Co. (the Boston scouting team, more or less) has been pushing for it as well.

Well I run my reliever as HGM suggests Torre should have (Use my best resources when the game is close and worse when I concede the game or I have a big lead) - but that is just a sim with numbers. Perfect numerical strategy doesn't always lead to the best results in the real world. "The best laid plans"...

[HoustonGM]

Yes but,as you know,the 2 situations are totally different - in the past (& it was argued quite expansively in another thread ) the SPs pitched deeper (& more often) & thus the "fireman's" role was largely different as he would come in & pitch "multiple innings" (close the game) but on a rarer basis

Not really a rarer basis. Look at Bruce Sutter. He routinely threw 60-70 games, which is about normal for a modern day closer, but just had more innings in those games.

Today there are far less CG's pitched by SP's & in general the Average innings pitched by a SP is down by 1 or 2 innings from the period you are "quoting".
Not to say I am for or against a return to a more traditional approach,but for it to be a viable solution the SPs have to also return to a "deeper" innings count.

There is a certain logic - best pitcher's used most often has been advocated for both by traditionalists & certain sabermetrician BUT as said above (& in the PITCH COUNT thread) this approach is only viable,today,if SP's pitch deeper than today or a 6 man rotation or whatever "system" which means RP's come into the game LESS often but are able to pitch further(or a return to 80's level).

You don't need to have your starters go further to more effectively use your relievers.

Well I run my reliever as HGM suggests Torre should have (Use my best resources when the game is close and worse when I concede the game or I have a big lead) - but that is just a sim with numbers. Perfect numerical strategy doesn't always lead to the best results in the real world. "The best laid plans"...

Tell me how in the world would throwing bad pitchers in tight situations in good pitchers in give-in situations would at all lead to the best results, over a large sample size like a season.

[WATERY]

who suggested over a season it would be better? Please reference and quote for me.

[HoustonGM]

who suggested over a season it would be better? Please reference and quote for me.

Well, that's what I've been talking about. I've made it clear a few times that while I made this thread because of one incident, it was because I feel that there is a much larger problem, with both Joe Torre and other managers, in that, over the course of a season, they fail to properly leverage their relievers.

[ltt]

all I'm saying is that it's not even mathematically sound to claim that managers should use pitchers differently than they do.

It is precisely the way Torre uses Rivera that produces the stats that are then used to insist he should use him otherwise.

[HoustonGM]

all I'm saying is that it's not even mathematically sound to claim that managers should use pitchers differently than they do.

There's tons of mathematical studies that show that managers can gain wins over the course of the season by better leveraging their relievers.

It is precisely the way Torre uses Rivera that produces the stats that are then used to insist he should use him otherwise.

??? Sorry, I guess I was under the assumption that the reason Rivera's stats were so good was because he was a good pitcher.

[ohms_law]

Itt doesn't think that statistics are real (or maybe either accurate or meaningful would be a more descriptive term), so it's rather pointless to argue about.

[FRENCHREDSOX]

Itt doesn't think that statistics are real (or maybe either accurate or meaningful would be a more descriptive term), so it's rather pointless to argue about.

Actually ltt is arguing the opposite Ohms.He is saying,mathematically,the stats produced are the ONLY real/accurate & meaningful (objective) because of the actual decision made by managers on player usage.

Whereas "leverage comparison" is mathematically incorrect/inaccurate & totally a subjective assessment because the situation :

1) happened under a different set (thus not comparable) of circumstances;
2) the numbers realized was a different match-up (because 2 pitchers cannot pitch to the same batter simultaneously );
3) there are NO real results.

Fact1 (from the post above) Torre USED Rivera to pitch.Fact 2 Rivera pitches & obtains WHATEVER result.Fact 3 The result DETERMINES Rivera's stats.In a "leverage comparison" you are comparing objective FACTS which are determined by the above order whereas a leverage comparison (of "what ifs" thus subjective & hypothetical totally) is mathematically impossible,thus disproving:

There's tons of mathematical studies that show that managers can gain wins over the course of the season by better leveraging their relievers.
.

Actually,There are tonnes of SUBJECTIVE statistical studies that show that managers COULD POSSIBLY gain wins ( or losses) over the course of the season by their usage of ALTERNATE POSSIBLE leveraging of their relievers


However, mathematically speaking the only objective results are from ACTUAL situation management (as described above - Torre lets Rivera pitch,Rivera pitches,result is ACHIEVED which produces the stats ONLY if Torre changes Rivera then a new set of unique objective stats /results are produced & again these cannot mathematically be disproved or approved just assessed in their relevance.)

[RickD]

Ow my head hurts!

[RickD]

BTW they won and clinched a playoff spot!

[ltt]

maybe I can clarify:

from my perspective, the argument HGM is offering looks like this:

A manager utilizes a pitcher a certain way.

This produces that pitchers stats.

Those stats are now used to argue that the manager should utilize the pitcher a different way.

to me, this argument fails in basic logic, before we even get to the philosophical disputes.

as to those philosophical disputes, I understand that HGM was only presenting this particular instance as an example of a larger general class of managerial actions that don't properly "leverage" relief use.

the problem is that the notion of "high leverage" situations improperly, imo, describes as a class a group of largely unique situations - that is, it treats all games tied in the top of the ninth inning as a "type" while ignoring the myriad factors that make each such situation different to varying degrees.

[ltt]

Itt doesn't think that statistics are real (or maybe either accurate or meaningful would be a more descriptive term), so it's rather pointless to argue about.

no, you've got it completely wrong. What I dispute is not whether statistics have meaning, but rather whether certain have the meaning you say they do.

I'm sure you'd like that to be a pointless argument, but since there seem to be passionate advocates on both sides, I find it highly unlikely that it should be so.

And surely you can do a better job of defending your position than just dismissing anyone who disagrees with you as ignorant of the holy truth statisticians have bestowed upon the world.

I mean, I could whip out my curriculum vitae and dismiss everyone I disagree with as wholly ignorant of the metamathematical thought that underpins my arguments, but I prefer trying to clarify my position and the flaws I see in others.

[FRENCHREDSOX]

maybe I can clarify:

from my perspective, the argument HGM is offering looks like this:

A manager utilizes a pitcher a certain way.

This produces that pitchers stats.

Those stats are now used to argue that the manager should utilize the pitcher a different way.

to me, this argument fails in basic logic, before we even get to the philosophical disputes.

as to those philosophical disputes, I understand that HGM was only presenting this particular instance as an example of a larger general class of managerial actions that don't properly "leverage" relief use.

the problem is that the notion of "high leverage" situations improperly, imo, describes as a class a group of largely unique situations - that is, it treats all games tied in the top of the ninth inning as a "type" while ignoring the myriad factors that make each such situation different to varying degrees.

The classic differentiation of a mathematical fact & a statistical hypothesis.(which is what ltt said)



HGM is arguing that a hypothesis will produce a better result which ,mathematically,cannot be true as the event HAS already occured.Thus the hypothesis,ceteris paribus,is POSSIBILITY of a different outcome but cannot be verified as the action is over thus it is actually simply a subjective theorem not a fact.

[Rongar]

The classic differentiation of a mathematical fact & a statistical hypothesis.(which is what ltt said)



HGM is arguing that a hypothesis will produce a better result which ,mathematically,cannot be true as the event HAS already occured.Thus the hypothesis,ceteris paribus,is POSSIBILITY of a different outcome but cannot be verified as the action is over thus it is actually simply a subjective theorem not a fact.

This is getting much too much cerebral for me - I knew that I should have stuck with mud wrestling! (notice my witty pun, Guys? - I'm cerebral too!)

[HoustonGM]

I'll just keep it simple.

The goal of each individual baseball game is to win. High leverage situations are, by definition, the most important times in a given game. What is the best way to maximize your chances of winning the game? By using your best players. You don't help your chances of winning by saving your best players just in case you're winning later in the game.

Yes, I understand, you can think of scenarios that explain why Torre wouldn't use Rivera in this situation. Fine. Whatever. This doesn't mean that, as a whole, managers are properly leveraging their relievers.

from my perspective, the argument HGM is offering looks like this:

A manager utilizes a pitcher a certain way.

This produces that pitchers stats.

Those stats are now used to argue that the manager should utilize the pitcher a different way.

to me, this argument fails in basic logic, before we even get to the philosophical disputes.

Throw stats out the window then. I'm arguing that the manager should use his best relief pitcher in the most important situations, and save his worst relievers for unimportant situations.

Take a look at what good leveraging of relievers can do. When does that manager use his best relievers? When the game's close. When does he use his worst relievers? When the game's not close.