Disclaimer: the following is of negligible analytical value. However, after a weekend at the Sloan Sports Analytics Conference and Ben Blatt’s lovely Six Degrees project, I realized that it was worth a post.
It was two New York summers ago — I don’t remember when, exactly, but my friends and I were watching the evening slate of baseball games. The Mets were probably losing. Those of us there had two things in common: boredom and a love for ’90s utility infielders, the very same combination that likely spawned PECOTA. We created something a bit more modest: a trivia game called Connection.
Here’s how it works: Person A starts out by naming a team in a particular year (e.g. “1995 Cleveland Indians”). Person B must respond by naming a player who recorded at least one appearance for that team, followed by a different team that player was on (“Manny Ramirez, 2010 White Sox”). The process repeats itself — if more than two people are playing, Person C would have to do the same for the 2010 Sox. If it’s just one-on-one, it goes back to Person A. You can only use players and team-year combos once per game, and you can’t use the franchise you’re given on that turn (for instance, you couldn’t say “Gordon Beckham, 2009 White Sox” in response).
Elimination works with strikes: three and you’re out, and the last player standing wins. Let’s say Person C responds incorrectly. It’s a dicey situation, but he/she might still be safe: Person B now has to make a connection from the 2010 Sox. If the attempt is correct, Person C gets a strike. But if not, Person B gets the strike. The next round starts off with whoever got the strike naming a new team-year combo.
We play Connection with MLB, but it works for any sport and league. If you and your friends know your stuff, this can go on for a long time. It’s a fun car game, provided that someone has a smartphone with access to the Sports Reference sites (most turns are usually easy enough that someone can confirm a right/wrong answer without looking it up). Here are a couple shorter, more horrifying variants for baseball.
- OPS+: Each player named has to have either an OPS+ or an ERA+ 100 or above (meaning they’re at least league-average) in both of the years involved. If he was on more than one team in that year, use his OPS+ with that team only. This is not as bad as you’d think: most players are memorable because they’re good.
- OPS-. Same as above, except the player has to have an OPS+ or an ERA+ 100 or below. NL pitchers’ hitting stats don’t count. This is really just awful, because even the worst hitters and pitchers seem to have one decent year. For instance, Jose Molina had an OPS+ of 101 with the 2007 Yankees. I learned that the hard way.
As I said above, there’s nothing too analytical here outside of the advanced stats used in the variations (I imagine you could use PER above and below 15 for the NBA. And maybe H-O-R-S-E instead of strikes). But after spending this weekend with like-minded people, I realized that the analytics community might enjoy the game. Let me know how you like it! Feel free to start a round in the comments.
Acknowledgements: Connection’s co-creators are my high school friends Sudesh Kalyanswamy, Ben Kleinman, and Ross Carstens. If you came up with and wrote about it previously, let me know and I’ll happily include you.