Carl Cervone
NYC Pizza Bracket: A Quest to Find The Best Pizzeria in New York City
July 19, 2014
Check out the complete results of the Pizza Bracket here.*
RULES
DESIGNING THE BRACKET
Inspired by Nate Silver’s Burrito Bracket, this is a NCAA-style knockout bracket to find the best pizzeria in New York City.
We took the boroughs of Manhattan and Brooklyn and divided them into four groups – West Manhattan, East Manhattan, West Brooklyn, East Brooklyn.
Each group has four pizzerias, for a total of 16 pizzerias across the bracket.
The bracket has three rounds:
- Round 1: we visit each pizzeria and choose the best for the group.
- Round 2: we go back to the winning pizzeria for each group and choose the best for the borough.
- Round 3: we go back to the winning pizzeria for each borough and choose the best pizzeria in the city.
Thus, a total of 16+4+2=22 pizza meals will be consumed over the course of the competition.
## CHOOSING PIZZERIAS
We set out to identify the four best pizzerias in each group. We created a shortlist from Yelp searches and then corroborated the list with reviews from PizzaMaking.com and Slice. A list of 22 eligible pizzerias was generated. We then devised a simple scoring formula and ranked the pizzerias based on Yelp reviews. For the most part, the Yelp formula created a ranking that was in line with our hypotheses about who should be on the list, but we had to intervene in a few cases.
-
Grimaldi’s was removed in favor of Juliana’s, which is next door. These restaurants are rivals with intertwined histories but similar styles of pizza.
-
Via Tribunali was a very competitive option, but omitted because its origins are on the West Coast.
-
Franny’s was selected as the fourth seed for West Brooklyn, even though Peppino’s had a higher raw score. We felt the majority of Franny’s low scores on Yelp came from naive customers who complained about pizzas not being sliced. From our perspective, Franny’s should get positive points for not slicing its pizzas.
-
Forcella has locations in Williamsburg and NoHo. We opted for the NoHo location (and the East Brooklyn group was already full).
-
Co. was selected over Otto as the fourth seed for West Manhattan, as we hold a slight bias against restaurants with celebrity chefs.
There are also some interesting trends that emerged from this analysis, which support the widely held belief that Brooklyn has better pizza than Manhattan.
- Average Yelp score: Brooklyn (4.1), Manhattan (3.9)
- Share of 5 star Yelp reviews: Brooklyn (43%), Manhattan (34%)
Overall, the most competitive group appears to be East Brooklyn, with 47% of Yelp reviews awarding 5 stars. The least competitive group is West Manhattan with only 31% 5 star reviews.
THE FINAL BRACKET
The scores were used to seed the pizzerias into the bracket.
SCORING THE PIZZAS
We created a consistent set of scoring rules for each pizzeria. At each restaurant, we would try a basic margherita and then one or more specialty pizzas (with toppings). The pizzas would be scored against five parameters:
- Crust
- Cheese
- Sauce
- Toppings
- X-factor (ambiance, service, price, overall vibe) Each parameter has a score of 1-10, for a maximum of 50 points per review.
Both Dan and I would tally up our scores and add them together, for a combined maximum of 100 points.
The pizzeria with the highest score in its group would advance to the next stage of the knockout competition.
RESULTS
West Brooklyn
Here are the summarized scores from the West Brooklyn group.
We also created a ranking of the different individual components. If a component had the same score, then we used standard deviation as a tie-breaker. A lower “stdev” means that Dan and I were more aligned in our assessment and tips the balance in favor of that pizzeria.
Crust seems to be a weak spot overall for West Brooklyn pizzerias. With a tie between Lucali and Franny’s, Franny’s got the win due to a lower standard deviation.
Lucali was a clear winner in the cheese, sauce and “x factor” categories.
Franny’s nailed the toppings category… I can still remember those ramps and dandelions.
West Manhattan
West Manhattan was the least competitive group heading into the bracket, with an average Yelp score of 3.9 and only 31% of reviews earning 5 stars. Well, the region definitely had a weak showing. But before we give Yelp too much credit for predicting this outcome, observe that the group’s top pizzeria (Keste – 80 points) and its worst performer (John’s – 35 points) had nearly identical Yelp ratings (4 stars average, of which Keste had 42% 5 stars and John’s had 41%). We can speculate the reasons for this but firmly stand by our smashing of John’s.
Keste also mopped up the individual categories (and John’s mopped the floor), with the exception of toppings, which went to Co.
East Manhattan
If we had an afternoon to tour New York pizzerias, we’d spend it in East Manhattan. This group had three excellent showings; any of the three could probably have won the group on its best day. Forcella clearly distinguished itself on the day we went and we are excited to return there for the finale to see if it’s the norm or a fluke.
Forcella was the hands-down, unanimous winner in the crust category. Luzzo’s delivered the best cheese AND the best sauce in the entire competition. Both Forcella and Motorino had outstanding toppings. (We are still craving Forcella’s lemon and burrata pizza.) None of these places had Brooklyn-level crowds, but Forcella’s extremely relaxed vibe and al fresco seating made it the clear favorite for the “x factor”.
East Brooklyn
We could probably do a rematch of the East Manhattan group in under four hours, from start to finish. The East Brooklyn group would require at least two full days of effort. Waits are long: you’d be lucky to get a seat at Roberta’s in under 2 hours (unless you have a reason to be in Bushwick mid-day on a Tuesday). We encountered serious crowds everywhere we went, except Barboncino (Sunday brunch). And distances are long.
Our pizza experiences were also all over the map.
Despite the, ahem, hardships involved, we found our two favorite pizza places in New York in East Brooklyn: Roberta’s and Paulie Gee’s. They tied at 89 points, the highest scores in the entire competition. But the tiebreaker, an anonymous vote from Dan, myself, and our partners, was unanimous in favor of Roberta’s.
Roberta’s and Paulie Gee’s were neck-and-neck in virtually every category. Paulie Gee’s had a one point advantage in crust, sauce and toppings. They tied in “x factor”, both with perfect 10s. However, Roberta’s won the cheese category decisively. And, in doing so, Roberta’s became the only pizzeria in the competition to score an 8 or higher in every category. Roberta’s sum was greater than her constituent parts, and this lifted the Bushwick institution into the finale.
Bracket Results
And here is the completed bracket for the Round of 16.
And here are some graphs for those of you who like pizza analytics.
First, we can see that Yelp scores had little relation to our scores. If anything, the correlation below is slightly negative.
)
From this analysis, we can then see which pizzerias are over-rated on Yelp. (We certainly were not impressed with John’s or Prince St.) Lucali is the only group winner that also appears on the hyped-up side of the line. Similarly, we observe that Franny’s and Forcella deserve a bit more credit on Yelp. But Paulie Gee’s is earned its 4.5 Yelp rating.
Last, we performed a simple test to see how aligned our scores were to each other. A high difference in points means we were not very aligned or, in other words, our individual ratings were less consistent. Sottocassa, our first pizza review, definitely stands out as a major outlier (and Dan confesses that his scores were a bit harsh). The trendline shows that our scores were more consistent for the better pizzerias. At Roberta’s, our favorite in the group stage, our scores differed by only 1 point. What you can take away from this is that we liked the same attributes in the good pizzas, but were less aligned in now much we disliked that weaker pizzas. Overall, Dan was a slightly harsher critic, with a mean score of 35.4 versus my mean of 36.0 (our median scores were both equal to 37).
