FiveThirtyEight: Don't be fooled by small-budget successes.

[weams]

That's all heart-warming, but evidence suggests that the relationship between money and winning is as strong now as it's been any time in the free-agency era. Check out the figure below, which shows the relationship between spending and win percentage during each of the three 10-year spans since 1985.1 Each team season is one dot in the figure, and the red line reflects a smoothed curve fit through the points. The smoothed curves represent the general relationship between spending and performance for each team season in each decade; aggregating all the decade's data points shows a pattern: More money generally means more wins.

link here

[weams]

Winning a playoff series can come down to a few factors — a couple of good pitchers and luck — that are less important during the regular season. “The formula seems to be: limp through regular season, get into playoffs, then win,”

[Frobby]

Nothing too surprising here. We all know that the ability to spend more money provides a cushion against stupid mistakes. But it also has been proven that the situation is not hopeless for the teams who are not at the top of the spending curve.

[DrungoHazewood]

Baseball made a pragmatic decision: they couldn't undo history, they couldn't go back and fix the massive disparities in revenue due to territorial rights and RSNs. So they expanded the part of the season where randomness rules - the playoffs. It's not the most elegant solution, but it mostly works. All you have to do is build a 85-ish win team on true talent, get a few things to break your way, and on October 1st you have almost the same odds of winning the Series as a 98-win, $250M superteam.

[Ohfan67]

Meh. That's not the greatest analysis in the world. It's actually a weak statistical argument. First, the data are pseudoreplicated. The "team year" method used means that rosters that stay relatively intact show up multiple times in the data set. I guess that's ok, but the '90's and early '00's Yankees team with multiple exceptional HOF's basically over influences the result. But that's not a big deal. It's the lack of a real statistical analysis that makes this piece "meh". The authors show a trend line, but they don't report an R-squarea and p-value. There's a ton of noise around that line, suggesting that the amount of variation in winning percentage explained by budget is actually very small. Furthermore, you can't even calculate a "correct" R-square unless you transform the winning percentage (actually expressed here as a proportion). Proportion range from zero to one and you would need to arcsine-square root transform the proportion to run the analysis and the graph of budget versus the transformed winning percentage (proportion) would be the accurate presentation of the data. I think this "analysis" is very weak.

[DrungoHazewood]

There's a ton of noise around that line, suggesting that the amount of variation in winning percentage explained by budget is actually very small.

Baseball probably has the smallest spread in winning percentage among major sports to begin with. If you think of MLB in NFL terms the worst teams go 5-11 or 6-10, and the best teams 10-6. Other sports routinely have teams that win or lose 80% of their games. So whatever the components are that contribute to variation in winning percentage they're going to relatively small in magnitude compared to other sports. Heck, just randomness explains probably half the variation in records between teams; a perfect projection system can't get any closer than +/- five wins or so.