Rounding Up the Guardians | Baseball Prospectus

Image credit: © David Richard-USA TODAY Sports

Every year, mathematically illiterate pundits mock the fact that projections report teams’ forecasted win totals to a precision of tenths of a win. “You can’t win half a game!” some wannabe Chris Russo invariably shouts. Sometimes, it’s even the original version doing it.

That’s silly, of course, because it’s reductionist. Projection systems like PECOTA aren’t actually suggesting that, for instance, the 2024 Guardians will win 83.4 games. They just achieve their robustness by way of running thousands of simulations, and reporting the results of those simulations to a tenth of a win is the best way to clearly inform consumers of the projections. It lets us know which way we’re rounding, if we do label a team based on the whole number of wins we might say they’re projected for, in shorthand. Every team’s record is made up of well over 81 games won and well over 81 lost; it’s just that a lot of those overlap. Some games are won, then undone by the failure of teammates. Once we admit that, it just makes sense to divvy wins up into win fragments and show everyone that work, or at least hint at it.

So, let’s talk about the 2024 Cleveland Guardians, because that example wasn’t chosen at random. There’s a good case to be made that the Guardians are the most round-uppable team in baseball this year. In fact, we might want to round them up by more than 0.6 wins—and at that point, it’s time to reconsider whether the Twins figure to have a clear edge in the AL Central this year.

Firstly, all projections are playing-time projections ($1 to Harry Pavlidis), and injuries are the greatest source of error in projection systems, so don’t overlook the vital findings published by Rob Mains and Derek Rhoads on Monday. We can’t assume that the Guardians will repeat their extraordinary achievement in the field of not losing time to injuries from last year, but if there’s any skill involved in that (as we have some reason to believe), it could be worth enough net win fragments to get them up to 84 wins, on its own.