Roberts Must Go

[RShack]

Yeah' date=' my dad got me hooked on some of his stuff. I would have to agree with the premise in this case, but then I also believe the acquisition of Bedard is going to help my team make the playoffs - go figure![/quote']

This is in no way inconsistant with the Foundation Trilogy. They play them one game at a time.

You guys are way more fun than Baltimoron is. (He's way too big on hate and hating, IMO.)

Plus, you're focusing more on Actual Baseball too.

My theory is that these two factoids are prolly related ;-)

[Rob H]

I don't do that about serious guys who are behind baseball analysis. Never.

I am "contrarian" (or correct, depending on your perspective) against some of the things *other people do* with their work... especially the inappropriate and unwarranted blanket conclusions that some people assert. Just because people look up stats on the internet, that doesn't mean they're Bill James... or Tom Tango. (Plus, sometimes even Bill James is wrong. I think the world of him, he's one for the ages, but still.) On a related note, I think highly of Freud and Jung, but I've known more-than-several highly-qualified Freudians and Jungians who were just-OK, and a few who were wacko to the point of being harmful. I've also known people who think they understand Jung because they read a couple books. Same kind of thing. If I challenge something that a book-club Jungian says, that doesn't mean I'm being contrarian about Jung.

When somebody jumps to unwarranted conclusions, I generally think it's completely innocent. Most of the time, I don't think anybody *means* to do anything inappropriate. I think the main problem is that reaching appropriate conclusions just requires some perspective, that's all. And, as I'm sure we can all agree, perspective is frequently lacking on message boards (in general, nothing specific to OH). The only time I think somebody's up to something is when they keep making the same mistake again and again and again, despite numerous conversations about it. And the only time I get really annoyed is when people start taking the position that I'm somehow up to something, or that my points are not germane. Or when they put phony words in my mouth.

I never said they were. Somebody said a million Monte Carlo simulations can't be wrong, and I said, "Sure they can". Because they can. I don't know all the details of what people do with Markov simulations. I would assume that a few people use them wisely and with great insight, while a few others are reckless with them, with many folks being sorta drably in between. I say this based on how people are in general, including researchers of all kinds. I would be very reluctant to jump to conclusions about whether the results are trustworthy, but this has a whole lot more to do with the great complexity of baseball than it does with any shortcomings of the researchers.

This might be simpler to grok if we just step back from baseball. I think it's dumb to trust the results of hurricane simulations for exactly the same reason. Are they worth doing? Yep. Does that kind of work contribute to furthering the state of the art? Yep? Are they gonna tell me where the hurricanes are gonna land next summer? Nope. I live where hurricanes come. Hurricanes are on my mind for half the year. I *like* hurricane researchers. I'm on their side. I also think Global Warming is real and we should do stuff about it. Yet whenever somebody tells me that Ivan and Katrina showed up at my house because of Global Warming, we have a little chat about how that's not exactly right. Same exact thing.

This all seems like a lot of effort because you didn't catch my allusion to the classic "50 million Elvis fans can't be wrong."

[RShack]

This all seems like a lot of effort because you didn't catch my allusion to the classic "50 million Elvis fans can't be wrong."

I hate it when that happens ;-)

ps: It wasn't you. It was the other guy.

[Rob H]

I hate it when that happens!

;-)

You never hated it in any of the Monte Carlo simulations I ran.

Edit:

Oh, and FYI. I do actually know what I'm doing. I realize you're just trying to keep everybody honest, but that's gonna clash real badly with my tendency to use the easiest method possible to prove my point, even if it's a bit innacurate.

[RShack]

You never hated it in any of the Monte Carlo simulations I ran.

That is funny. Respect.

[Baltimoron]

The central premise of the Foundation Trilogy is that human events can be predicted with some degree of accuracy in the aggregate, but that the actions of any individual can not be so predicted. It anticipates chaos theory. This paradox is the crux of the argument between Baltomoron and RShack IMO.

Yes, you brought this up with me regarding this similar issue a year or two ago and we had a back and forth.

This has absolutely nothing to do with the issue here, just as it did last time (the context was projection systems).

The issue is the love of beating up strawmen. Its lazy, misleading and ignorant.

[Baltimoron]

I never said they were. Somebody said a million Monte Carlo simulations can't be wrong, and I said, "Sure they can". Because they can. I don't know all the details of what people do with Markov simulations. I would assume that a few people use them wisely and with great insight, while a few others are reckless with them, with many folks being sorta drably in between. I say this based on how people are in general, including researchers of all kinds. I would be very reluctant to jump to conclusions about whether the results are trustworthy, but this has a whole lot more to do with the great complexity of baseball than it does with any shortcomings of the researchers.

He was right. The monte carlo simulation is not wrong in this sense (unless the math is wrong), the error is the human who is (arguably) incorrectly interprets and/or extrapolates from the model.

You know nothing of the model you are referring to and seemed to innocently and fairly be talking about markov sims and their potential shortcomings, but then you jump into your nonesense in paragraph 2 about what is silly and what you should be wary of without regard to the fact you don't even know about the model and methodology you are criticizing.

I'm not sure what you're citing them as evidence of, but many simulation techniques imply at least a couple basic assumptions that IMO are highly dubious when it comes to baseball: (1) that any-and-all attributes of the problem-space which aren't reflected in baseball stats are either nonexistent or have no value, and (2) that serial events in a ballgame (and a season) are discrete events that are unrelated to each other. In lieu of knowledge about such things, simulations techniques typically rely on simply inserting randomness.

Personally, I think that's a silly thing to do. However, evidently there is some difference of opinion about this. If you believe that stats capture everything that matters, and if you believe that baseball events are truly discrete events, then you might trust the simulations. On the other hand, if you believe stats don't capture everything that matters, and/or if you think that serial baseball-events are (or can be) somehow related and intertwined, then you prolly should be very suspicious of them.

[Number5]

Indeed. I'm just trying to point out to those who think that losing Roberts will leave a gaping hole at the top of the order should take a sec and realize that getting on base is what's important; not OBP AND speed.

I have to respectfully disagree. OBP is certainly the number one trait desired in a lead-off hitter. It is hard to score if you aren't on base. However, speed is the number two desired trait. The ideal lead-off man can get himself into scoring position for the 2-3-4 guys and needs to be able to score from second on a single. Yes, it is true that the Red Sox do well with Pedroia and Youkilis at the top of the order with high OBP's and few steals, but that is more a function of having Ortiz and Ramirez hitting 3-4 don't ya think? Given the choice between the two, yes OBP is more important than speed, but that does not mean that speed is unimportant.

[Bloopandablast]

Roberts was a .285 EqA hitter last season. DeRosa was .268.

Roberts is projected for a .293 EqA next season. DeRosa is projected for a .270.

Assuming an equal amount of PA for each (let's call it 600... split the difference between Roberts likely getting 650 batting at the top of the order, and DeRosa getting 550 batting lower)...

If they both perform like last year, Roberts is worth about an extra 11.5 runs. If they both perform to their projections, Roberts is worth about 16.5 extra runs. That includes stolen bases, but it doesn't account for the rest of their baserunning or their defense.

All in all, Roberts is probably about a 25 run upgrade over DeRosa.

Help me understand the basis for the difference in EqA last year when OPS was very close. What factor would be weighting it causing it to be more than marginally different.

The answer to that is BB's are weighted at 1.5 in the EqA vs in OPS and the fact that stolen bases are included in EqA. Since I struggle with why a BB has more weighting that a hit (particularly when a hit has a better chance or driving in a run) I would suggest that the only real difference is back to my original point and that is the biggest thing you get out of the deal is base running and maybe some defensive range.

Others will then use the "synergy" arguments: 1) Soriano is not in the leadoff spot and therefore will contribute more 2) DeRosa is allowed to be the supersub and therefore depth is better 3) Better left right balance in the lineup since he is a switch hitter.

But numbers 1 and 3 above are really dangerous reasons to give as true advantages of the deal. The balance already plays itself out in the OPS so that would be double counting. Who knows how Soriano will react. He may struggle in the 4 or 5 spot. Depth is improved, at least in flexibility.

[Rob H]

Help me understand the basis for the difference in EqA last year when OPS was very close. What factor would be weighting it causing it to be more than marginally different.

Adjustments for league, park effects, and team pitching can have some decent effects.