Monday, June 13, 2005

Ok, now it is getting eerie

You might remember a few weeks ago, I posted asking the question: at what point can a fan become statistically significant? And I have kept y'all updated here and there on the Nats record when at least one member of Nats Blog attends. Well, the convergence of us going to a bunch of games over the past two weeks and the Nats going 12-1 on this homestand (none of us attended the one loss), has put the Nats' Nats Blog attends record at a stunning, preposterous 17-0. Yeah, that isn't a typo and I am knocking on wood like crazy (ok, well, pressed particle board, I don't think this is genuine wood).

That means, the Nats are 17-0 when we attend and 7-9 when we don't. What are the chances of that? Well, since you asked, about 1 in 3300 (for math geeks: the chances of randomly having all 17 games you attend result in a win when a team is 24-9 overall is 24 factorial up through 8 divided by 33 factorial up through 17, which equals about a 0.0003 chance).

Now, the question I'd have to run some real analysis on, and if anyone with real skillz out there feels like doing it is: Can one actually say that based on the above, Nats Blog attendance is statistically significant, and to what degree of confidence. The issues that are twisting a bit in my head are whether for the sake of analysis, it matters that I don't know how many "sets" of season ticket owners there are (if 22,000 season tix, maybe 8,000 sets of owners?), and I haven't thought about whether it matters that I don't know how many games those season ticket owners go to.

Anyway, it would be interesting to see and fun to be able to say. Now, of course, we KNOW that this is absurd and are not so full of ourselves that we think we truly affect the outcome (ok, maybe I do), but there are things that happen all the time that we know have no correlation and are just absurd outlier coincidences, but are nevertheless statistically significant when you run the math (e.g. before this past year, Presidential elections based on whether the Redskins win their last home game, etc.). And besides, 17-0 with us there and 7-9 without is a pretty cool stat and I'm going to milk it for now!


At 10:37 AM, Blogger Carl said...

I'm 8-1 at home games this year. Not too shabby. The only loss was the 3-0 Loaiza-Myers pitching duel, and I'm a lifelong Phils fan. Go figure.

At 11:43 PM, Blogger Scott said...

So -- I indirectly asked my college roommate about your freaky win streak (meaning -- I linked to it at my place and he picked up on the hint). He's a psych prof, so he does statistical mumbo jumbo all day -- I don't know a chi squared from a T square. Anyway -- here's what he found:

The chisquare for Natsblog is 13.15 with one degree of freedom. My statistics tables only go to p = .005 and that value is only chisquare of 7.88, so this is significant to an extreme degree. If this were a scientific study, you could publish it in Nature.

Serious mojo, people. Y'all appear to be tuned into the deep wells of the baseball fates. Cool.

Love the site. Keep it up!

At 11:48 PM, Blogger DM said...

Yikes! I'm getting on a plane to Anaheim. I mean ... Los Angeles... or California.

Maybe Bowden will pay us $4.25 million. :)

Thanks scott and your college roomy.


Post a Comment

<< Home