Bridging The Great Baseball Analysis Divide

is a great tool to look into as well – http://saberlibrary.com/

There’s been many many intro pieces done recently, big league stew (Yahoo), lookout landing, etc. etc.

What needs to be done really is for a dialogue to be struck up with those more in the mainstream, such as GKR, to get them to understand and to potentially use such analytical tools. The use of OPS has become mainstream, but the key to me is that it is mentioned on the air in broadcasts.

Getting into a dialogue with broadcasters such as GKR about BABIP and FIP, which aren’t too complicated, is the first step. (wOBA might be another thing to bring along, as might WAR, but linear weights tend to be confusing for people.)

I don’t see an intro to stats here being particularly useful, as it’s essentially as you just said, preaching to the choir. Now if metsblog ran such an intro it might be better, and if GKR tried to explain a concept and use it in broadcasts, it’s be effective. But right now you aren’t bridging as much as preaching to the choir.

… when they often seem to contradict what seems obvious to the naked eye. For instance, last year’s stat (I forget which one) which suggested that Daniel Murphy was the best fielding 1B in the league in ‘09. That just ain’t right. Also, coming into ’10, I had burned into my noggin that Bay is a disaster in LF, which clearly is not the case.

I know in his early work Bill James was never comfortable with the statistical means of evaluating fielding. Is UZR really that valuable, and how valuable can it be when different people’s versions come to wildly different conclusions?

I’m not being a luddite smartass, btw. I really want to know.

An aside: one doesn’t need to be conversant with the more complex saberstats to know Francouer is a mediocre-to-poor baseball player. A mere knowledge of so mundane a stat as OPS tells us that … which tells me that Mets management is, at best, toiling in the Bronze Age of statistical analysis.

I think it would be valuable to dicsuss what advanced stats can tell you about the future. For example, Ryan Howard’s big deal with the Phillies—the raw numbers he piled up made it seem like he was worth every penny, but more advanced stats betrayed tendencies that would make Howard less valuable in the future. (We could tell the same story about Castillo, I suppose.)

Since many people have been compiling stats for a while, it may be instructive to look at players in 2006, review what Conventional Wisdom said versus what stats predicted, then see how things panned out.

May I request some work with wRAA? I’m much more comfortable with wOBA, but I think a counting statistic is one that might be more acceptable to the less sabermetrically inclined.

The folks that need to read won’t and those that care already have read about the Sabre stats.

I was listening to Gary inane rant the other day that you wrote about and it’s clear to me he doesn’t give a damn about tying to understand. He wanted to say David Wright is awesome and he’s above reproach and didn’t care what he had to trash to it. Our SNY team IMO has read to many of their press clippings and have become guilty of the same bias and nonsense the rest the press in this town is guilty of of.

I am a sometimes, if rare, poster, mostly because I do not understand what is being reported. A key would even help. When you get blood tests done at the hospital, the results come back with a normal range, #>#. Something along those lines would be a great help. Such as, a players UZR should be X to X if they are good or X to X if they are mediocre. That is how I learned OPS, ERA, BA, Slug%. I would compare say Piazza to the rest of the league. Too bad more sites do not list the sabermetrics, so I can compare them more. Overall, love the idea!

Will this series complete before or after the Top 50 Mets of All Time?

When explaining a stat, include the strengths and weaknesses of the stat (and perhaps link to other stats that address the weaknesses to an extent).

Speak their language, just say different things.

Before you write, maybe have another look at the ‘Introduction to Player Ratings" chapter of the New Bill James Historical Baseball Abstract, where he lays out the concepts of “Runs Created” and “The Win Shares System” in plain, patient English. If I had never read this, I probably wouldn’t give a shit about baseball statistics to this day. I can vividly remember the revelation that was its first few paragraphs:

Well then, how do we know how many runs a player has created?
There is a very simple formula which will answer that question quite well 90% of the time. The formula is:

Hits Plus Walks
Times Total Bases
Divided by (At Bats Plus Walks)

If you run this formula for any team since 1920, it will almost always give you a total within 5% of the team’s actual runs scored. In 1999 every major league team scored within 5% of the number of runs projected by this formula.

Wow!!!

This was the hook: numbers can tell you amazing things about the game of baseball.

I think a detailed explanation of metrics would be helpful (For example, why is the value of an error more than a single in wOBA).
Also links to sites that have done research, like Pizza Cutter’s research into sample size.

I’m glad you’re doing this. I’m one of the three, or five, or seven (or ten!) that want to better understand this stuff, so thanks in advance!

Thank you for doing this. I look forward to being reprogrammed for the 21st century.

As someone fairly new to sabermetrics I appreciate posts likes this. Thank you

I have too many other things in my life to start learning all these new terms and stats…it is the same reason I do not buy one of these new cell phones that do everything but wipe your butt. Sure I would like to learn but I have other priorities to tend to.

I think the big thing would be, when talking advance stats just put the full name of the stat first. like I didn’t know BABIP was, but right in the first line you explained it. now I don’t have to look it up. I was always taught, when writing and using acronyms, to always put what it means the first time you use it in a story so the audience knows what you are talking about. Especially when we all have different jobs, and some of the acronyms I use at work everyday for something totally different

Sometimes a new voice or new phrasing can be helpful, even if you’re already familiar with an idea. As to specific subjects, I’m always interested in hearing more about the “What influence do pitchers have over batted balls?” question.

and then a series of posts doing a deep dive into each, but without the “my stat is better than your stat” tone that often creeps into these discussions.

When I came here, I was skeptical. Over time, I’ve come to appreciate those stats that I understand and make sense to me. I am still waiting to be convinced that ALL the saber-stats are useful. Some of that is due to the fact that often, when I ask a question, I get conflicting responses.

Until then, I will continue to hum “WAR – was is it good for – absolutelty nothing – say it again” everytime I open up a thread about stats.

I may be in the minority here in terms of what I want to understand about some of these stats that I don’t. (I’ve been experimenting a bit on my own with teaching myself some of these) I want to know the calculations. I want to know the logic and reasoning and why I’m trusting that when a guy analyzes something as a Line Drive over a Fly Ball that I can trust.

Conceptually it’s not a huge stretch to understand the meaning of say WAR, but I think part of the problem is the assumptions it makes. What are we defining as a replacement player, what exactly is a Win, and how much does it mean over what sample size? How accurate is it that an identical team with say Angel Pagan in CF over Nyger Morgan is going to win exactly the difference in wins that they represent?

And then, how do these stats relate back to Runs (The most important stat, in terms of measurement). Because that’s what we’re all ultimately after. How much is a given player contributing to our team scoring the most runs.