Sabermetrics are a bunch of malarkey . . .

[Russ]

Theoretical vs. actual. But neither side tells the whole story even if one side tells you a little bit more.

[Qwerty75]

Originally Posted by ctorg

I've been struggling to find a way to explain the viewpoint of people who go for traditional stats to others like me who go more for sabermetric measures. I have a sort of intuitive understanding of it that I can't seem to put into words correctly. That is, I understand it on some level, but my attempts to communicate that understanding fall short.

I think it's because sabermetrics add a layer of statistical analysis at a level the casual fan or those most familiar with traditional stats aren't generally equipped to, or care to, understand. For example, they'll look at .52 per stolen base and .26 per walk in the RC formula and say, "Where do those numbers come from? I'm not sure I trust/buy into this." The conventional stats have a long tradition of encoded value, e.g. .300 batting average is a benchmark, as are 30 HRs, 100 RBI, etc. Sabermetric stats take numbers that traditional fans feel close to and chop them up and regurgitate them into something foreign. .300/30/100 means something to the traditional fan, while 50 VORP has no mythical or emotional associations.

I didn't know what regression analysis or statistical correlation was before I began to investigate sabermetrics, so I can see where the unwillingness to accept the new stats at face value come from. However, I also think that fans that chain themselves solely to traditional stats are doing themselves a disservice by not inquiring into the new stats thoroughly and seeing what they have to offer. At least seek to understand the methodology so you can take the next step whether to assign value to or dismiss what researchers come up with. And to claim you have understanding of how they work while at the same time displaying evidence of the contrary is a disservice to those who seek clarity and depth in knowledge about baseball, which, at root, is just what sabermetrics strives to accomplish.

[RonCo]

Originally Posted by ctorg

I've been struggling to find a way to explain the viewpoint of people who go for traditional stats to others like me who go more for sabermetric measures. I have a sort of intuitive understanding of it that I can't seem to put into words correctly. That is, I understand it on some level, but my attempts to communicate that understanding fall short.

It's something like this:

Sabermetric stats will give you how well a player put himself in a position to add to his team's chance of helping the team win. It eliminates the luck of what his teammates did and all that and gives you a purer idea of a player's skill apart from happenstance.

Traditional stats tell you, factually, how often a guy drove in runs. They make no attempt at getting rid of luck or anything else. You just get what he did, rather than a more abstract theoretical value.

Now, the reason why I'm not satisfied with this is, well, all RBIs tell you is how many runners a guy drove in. The utility of knowing that is, well, questionable. Knowing how many players a guy drove in is kind of useless as far as determining how good he is. I mean, Steve Jeltz could have driven in 100 if he'd had a team of amazing players surrounding him and getting on base in front of him all the time. Would that make him better than a guy who hits 30 homers and walks a lot but only drives in 80?

So I always end up just turning my own argument around and making it about why sabermetric measurements are better, yet somewhere inside I do understand the desire to look to traditional stats instead. I just can't explain it.

There are no "sabermetric" stats and "traditional" stats. There are only data and different ways of configuring the data. Each way of configuring data provides information content of a quality that is dependent upon the question being asked.

The number of home runs a guy hits is a piece of data that we compile into a stand-alone stat called Home Runs. It's simple and quick and people understand it. It has no context, but us humans will always provide that all by ourselves. HR provides us with a measure of a player's power and his opportunity. See...there's a shot at minimal context. If a guy hits 55 HR, we can infer that he got a full season of plate appearances (opportunity)...but saying HR count is a good measure of opportunity or power is dangerous. I mean, what about a guy who hits 10? To understand if this is a measure of power or opportunity, we need more information content in our stats. Specifically in this case we need AB and ballpark as a minimum. Similarly, if we take that same piece of data (HR hit) and compile it differently--say multiply it by 4 and add it to 3B*3 and 2B*2 and 1B--we get a different stat with a different information content.

You use stats of any kind to draw conclusions about questions we ask. Like "Who was the best player in 2007?" If you measure players by a stat called "Stolen bases" you get one answer. If you measure it by a stat called "ERA" you get a different one. Baseball is about creating runs for your side, and stopping runs for the other side, all combining to create wins. So to answer this kind of question really requires a measure that relates closely to runs and wins. RBI certainly relates to runs...as does, well, runs. Seems simple to go from there to "Most Valuable" or "Best."

Add to this that humans have a dislike for complexity. The mere fact that lots of people say they like complexity (which generally provides higher quality understanding), but then show lack of patience to really understand the nuances of issues is an ironic measure of this very dislike for complexity. Don't believe me? Listen to politicians, or try to make a complex business presentation to senior management. Baseball fans are no different. It's easier to grasp simple measures. Besides, baseball in particular is all about the mystery of numbers. 56. 61. 73. 714. .406. My guess is you all know what those 5 numbers are even though three of them are no longer records.

But at the end of the day, those numbers, like every other statistic, are just numbers with varying levels of information content. The information content of any stat--let's say VORP--is only valuable if it answers a question better than the information content of another stat--let's call that one RBI. If you want to know the answer to "How many times did runners score as a result of Juan Pierre's plate appearances?" You can only get that by using RBI...of course. If you want to answer the question "How valuable was Juan Pierre?" and you have only RBI at your disposal, you get a cruddy answer.

You used the term "utility" above. That's a key point. Baseball fans should, to my way of thinking, be interested in getting the "right" answer to any question asked. Statistics...all statistics, regardless of when they were created or named...tell us these answers. The challenge is to use the right statistic to answer the question on the table.

The pages of this thread were created when a poster suggested that VORP should never be used to give the MVP award to a guy who had less RBI than a guy with a lesser VORP. Taken by themselves, the utility of VORP is greater for deciding this question than RBI. The reason for this is that VORP (and OBP SLG and OPS and several other stats) correlates to runs and wins at a higher rate and with greater significance than RBI. This is a mathematical fact that is borne out in study after study after study.

Baseball fans cannot have it both ways. If you understand this complexity, you understand why these measures are better for addressing the question of who should be the season-long MVP. But change the question and you change the tool that provides the best utility. Ask "Who is the strongest hitter in the league?" and I'm willing to bet the stat "Average Distance of HR" is a better identifier of that answer than VORP. Ask the question "Who is the Player of the Game?" and you need a different lens altogether.

Bottom line: Complex stats (I like that term better than Saber-stats) make people work harder to answer questions than they want to work. If simple stats get them 70% of the way there, a majority of folks say "That's good enough for me."

But for those who really, really want to know the answer...whose blood burns for deep understanding and who find their own mythical sense of power from truths that are more hidden...for those the simple 70% solution isn't good enough.

...well, that was fun.

[RonCo]

Originally Posted by Russ

Theoretical vs. actual. But neither side tells the whole story even if one side tells you a little bit more.

There are no two sides. There is only varying levels of truth and statistical confidence.

[endgame]

But for those who really, really want to know the answer...whose blood burns for deep understanding and who find their own mythical sense of power from truths that are more hidden...for those the simple 70% solution isn't good enough.

Classic, and without judgment. Excellent summary RonCo.

[Left-handed Badger]

Originally Posted by BoofBonser26

You do realize that getting guys on base, and around them, is kinda like the entire point of the game since that's how you score, right?

Yep, and you do realize not getting the runners in, generally make it hard to win. You know especially if the other teams has more runs at the end of the game than you do.

[wlight1]

Originally Posted by ctorg

I have a sort of intuitive understanding of it that I can't seem to put into words correctly. That is, I understand it on some level, but my attempts to communicate that understanding fall short.

Don't worry about it. Jestre can't put it into words correctly either.

RonCo, brilliant explanation of the issues. At the end of the day, what stats a person likes really does come down to the two questions: What, exactly, would you like to describe, and How well would you like to describe it. I think that personal takes on what "describing" constitutes are what's really at the heart of the "traditional" vs "sabermetric" or classically termed "stats vs scouts" debate.

[rasnell]

Originally Posted by RonCo

There are no "sabermetric" stats and "traditional" stats. There are only data and different ways of configuring the data. Each way of configuring data provides information content of a quality that is dependent upon the question being asked.


...well, that was fun.

Very good analysis, as usual, from our chart master and stat meister. Loved every word of it.

[Jestre]

Originally Posted by wlight1

Don't worry about it. Jestre can't put it into words correctly either.

Yup, I ar too dum to unnerstand vorp.... i thank rbis is da bess way to rait playa's jess becuz

[ctorg]

Originally Posted by RonCo

There are no "sabermetric" stats and "traditional" stats. There are only data and different ways of configuring the data. Each way of configuring data provides information content of a quality that is dependent upon the question being asked...

Really great post. It deserves to have some sort of star next to it or something.

[Russ]

Originally Posted by Jestre

Yup, I ar too dum to unnerstand vorp.... i thank rbis is da bess way to rait playa's jess becuz

roffle

[BoofBonser26]

Originally Posted by Left-handed Badger

Yep, and you do realize not getting the runners in, generally make it hard to win. You know especially if the other teams has more runs at the end of the game than you do.

Can't strand runners if they're not on base!

[USN Squid]

RonCo is wicked smaht.

[Russ]

true, shipmate

[USN Squid]

Originally Posted by Russ

true, shipmate

Flashbacks! Flashbacks!

I've been out for 4 years.

[Left-handed Badger]

Well, I'm done with this post

To all VORPers: Promise me you wont be on Oprah, jumping on her couch, professing your love to Keith Woolner.


And the lefty Badger is off back to the world of sanity.

[Boom Boom]

Sorry to dredge up an old post.
Over a course of a season, a player who hits 4 for 4 with 2 doubles, 2 singles with no runs scored, is more valuable than a player who only hits 1-5 with a grand slam.

RBI is team dependent. RBI and RC is to judge how well a team does. Not how well an individual player does.

If you want to look at an indivdual player, don't look at RBI, look at OPS, OBA, TB, SB &SB%, OPS+ and sluggling and some other stats.

as for VORP, it doesn't make sense to me, so I left that one out...but for RBIs, to look at a player who drove in 120RBI's or 100RBI's doesn't mean anything but how good each teams baserunners who hit in front of them were.

For example, who is better:
Player A .300avg, 30HRS, 100RBI (OPS: .999)
Player B .300avg, 30HRS, 120RBI (OPS: .999)

They are basically the same player, based off of those stats. Anything do to with runs scored, and runs against is total Team dependent. Any stat based off of runs, is silly to say that one player is better than other...You have to look at how that player performed for that said team. Runs stats: R, RBI, ERA, DICE...ect do not really say how well that player was at driving in runs. The runners would have to be there in the first place to have a high RBI total. I hope this makes sense.

[johnnybourbon]

Originally Posted by Boom Boom

For example, who is better:
Player A .300avg, 30HRS, 100RBI (OPS: .999)
Player B .300avg, 30HRS, 120RBI (OPS: .999)

Thats an easy one. All else equal, I'd take the dude with 120 RBIs. There's something to be said of batting average with runners in scoring position. sure, there needs to be people on base...but what good is a .OBP-.500 guy that hits below .100 with runners in scoring position? (exaggeration, i know, but I hope you get my point)

[BigCity]

Wow... can't tell you how many times this discussion has been on the board

After reading throu all the opinions, let me put the arguement in the simplist terms possible.....

I come to bat in the 1st inning, 2 outs, no one on, and I hit a homerun - 1 RBI.

I come to bat again in the 4th inning, 2 outs, bases loaded, and I hit a homerun - 4 RBI's.

When you analyze my two at bats, I did exactly the same thing both times, the 4 RBI's I got on the second at bat have NOTHING to do with my ability - nor does it suggest I was any better in the 4th inning than I was in the 1st.

RBI's are nothing more than luck of the draw - and if anything, are a better measurement of the quality of the team I played with.

Interesting huh?

ok, going back into my hole...

[Neags23]

Originally Posted by johnnybourbon

Thats an easy one. All else equal, I'd take the dude with 120 RBIs. There's something to be said of batting average with runners in scoring position. sure, there needs to be people on base...but what good is a .OBP-.500 guy that hits below .100 with runners in scoring position? (exaggeration, i know, but I hope you get my point)

How do you know the guy with 120 RBIs didn't come up to bat with 1000 more men on base? Maybe he was the one that hit .095 with runners on base and still managed to drive in 120. Perhaps the guy with 100 RBIs hit 1.000 with runners on base.

Might be that you just took the wrong player because he had more RBIs.