This feels kind of lame.
When I started the Burrito Bracket a couple of months ago, I declared the following about the judging criteria:
"[T]he scoring will be determined almost exclusively based on the taste and quality of the food; things like service, "atmosphere" and price will only be used as tiebreakers".
Since then, I've altered my focus a bit, writing extensively about the non-food portions of the burrito experience, and introducing the Burrito Breakdown, which attempts to put things into slightly more scientific context.
The bottom line is that the way I'm striving to evaluate this contest is something like this. A stranger walks up to me on the street. He tells me he wants a burrito. He tells me he's trying to decide between two places, and they're exactly the same distance away. He asks me for a recommendation.
If that stranger walked up to me today on the corner of Division and Wood, and told me that he was trying to decide between Tecalitlan [review] and Carnicerias Guanajuato [review], I'd hesitate for a bit, and tell him to go to Tecalitlan.
Now suppose I knew this person a little better. Suppose this person was a friend of my mom's, who hadn't eaten at a Mexican restaurant apart from Chi-Chi's in the past five years. I'd tell her to go to Tecalitlan without hesitation. Suppose instead that it was my burritophile friend who didn't give a shit about service. I'd tell him to go to Guanajuato. But by the "passerby test"? The reasonable man test? It would be Tecalitlan by a nose. The food might be a little better at Guanajuato, but just not quite enough to outweigh Tecalitlan's other advantages.
I do think, however, that the Burrito Breakdown needs to be brought slightly more into proportion with this subjective sort of judgment. Therefore, I am introducing a new category which slants things slightly more toward the food side. This category is called "Construction/Balance", and it deals with how well the burrito is put together. Maybe all the ingredients themselves are pretty good, but they're not put together in the right proportion? Maybe the way that the burrito is structured, it's inherently going to be pretty messy? Maybe one burrito is more than the sum of its parts and another one is less so? That sort of thing goes into Construction/Balance. And it's sort of an important category here, in that it was something of a problem at Tecalitlan; their burrito was overstuffed with meat, and harder to consume as a result.
Add all that together, and Guanajato gains some ground, but it still loses to Tecalitlan by half a chili pepper, confirming my intuition from the passerby test:
You'll see, by the way, that I've introduced half-chili scores to provide for a bit more discretion in my rating formula. I've also assigned letter grades -- if you're really curious about how these are derived, please see the comments section.
Anyway, Tecalitlan stays alive -- by the thinnest of margins. It's going to have it's work cut out for it if it wants to beat Irazu/Lobos in Round 2.
This feels kind of lame.
Posted by Nate Silver at 1:26 PM