THE BOOK--Playing The Percentages In Baseball

How much could selection bias affect the results? I know that I am much more likely to leave a negative review than a positive one, though I happen to be a rather negative, cynical person on average.

I’ve been to #s 10, 12, 13, 19, 24, 25, 29, 30.

The retired parks on this list I’ve experienced (Yankee, Shea) would rank 21 and 27.

I’ve got to do a better job of getting to quality parks!

Average is 4.027, my average is 3.761 for a Stadium Experience SE+ of 93.

Nate Silver has never produced or referenced a ranking as accurate as the one that puts Rogers Centre dead last among MLB ballparks.

I was at PNC Park ten days ago as part of a trip to go to games in Pittsburgh, Cleveland, and Detroit.  The reason for going was to see games played on grass in newer parks - as my only experiences have been at Toronto (including Exhibition Stadium).  The rankings of PNC Park and the Rogers Centre relative to one another are right on the money.

Just imagine how good the experience at PNC would be if the Pirates were once again in a pennant race!

I was surprised Nationals Park rated that low. I go to several games a year there and enjoy the park.

Not only is PNC #1 by average, but it has the lowest standard deviation of any of them. That means people consistently think it’s awesome.

Also top 6 highest standard deviations are Los Angeles, Oakland, Chicago, Chicago, Los Angeles, New York. I wonder if ballparks in big cities (and diverse populations?) lead to larger swings in ratings.

@6 - this is a necessary consequence of a limited ratings scale.  The only way PNC can have such a high rating is with a lower standard deviation. If it were rated a 5, the SD would be 0.  The ratings for the other parks you mentioned are generally much lower and in the meatier part of the bell curve.  The unexpected result would be a park with a mediocre rating and a low standard deviation.  For example, having a lot of 4s and very few 3s and 5s.

This is definitely a useful starting point, but you have a big problem because it’s not the same folks going to the different parks.
For example, people in Pittsburgh could just have a habit of rating everything they experience in life as either a 5 (=good) or 4 (=bad) while New Yorkers may reserve 5s for only the rarest of experiences, like a meal at Per Se, or shopping at Zabars, etc.
One way to start to get around that is to look at z-scores by person—so, how does the ballpark differ from the mean/SD that these people are otherwise using. But that still has the problem that Pittsburghians are comparing their ballpark experience to ... not sure exactly… while New Yorkers are comparing it to MSG, the Met, etc.

What would be better is a survey where you get ratings from people who have been to multiple ballparks and z-score that.

mettle: I agree. I would do it exactly like here:

http://www.tangotiger.net/movies/

I’ve been meaning to do just that actually…

I’ve been to 1, 4, 7, 9, 10, 11, 12, 14, 16, 18, 22, 23, and the Metrodome.

I personally love Kaufmann, I’m not from KC, and would put that as my number two (behind PNC).