What It Means
A hundred and twenty-eight seasons, three thousand games, and the questions Purdue fans have argued about for as long as there has been a Mackey Arena to argue in. Here is what the numbers say, and where they stop.
The greatest Purdue team
It is the one you just watched. The 2023-24 Boilermakers rate 2.53 standard deviations above the Division I field, second in the country that season behind the champion that beat them, and no Purdue team since full-field records begin in 1985 comes within a quarter of a standard deviation. On accomplishment, the thing a team actually won weighted by how hard it was to win, the same team edges Rick Mount's 1969 national runner-up by a few points, and the two of them stand well clear of everyone else. Those are the two Purdue teams that played on the last Monday of the season, and the numbers agree with the memory.
The surprise is what comes next. Gene Keady's 1987-88 team, which lost in the Sweet Sixteen, rates second by strength since 1985, ahead of Glenn Robinson's 1994 Elite Eight team and both of the Carsen Edwards teams. Twenty-nine wins, four losses, an outright Big Ten title, and a schedule the field ratings say was harder than it looked. The tournament loss to Kansas State is the reason nobody puts that team first. The ratings say it may have been the best team Keady ever had.
Before 1985 the field data thins out, so the site measures those teams only against their conference, which every opponent's full standings let us do cleanly, and labels the result coarse. On that measure the best Purdue teams are Ward Lambert's: 1933-34 first, then the unbeaten 1911-12 team, 1919-20, and, fourth, Mount's 1968-69 team. That measure agrees with the full ratings at 0.85 in the forty-one seasons where both can be computed, which is good enough to trust the order and not good enough to trust the third decimal. The site never ranks a coarse-tier season against a full-tier season without saying so.
The Michigan game
This project began with a memory: Robinson's 1994 team going into Ann Arbor late in the season and winning, against a Michigan team stacked with future professionals. The data has the game. On March 6, 1994, at Michigan, Purdue won by 1 against an expected margin of -0.8, over an opponent that finished the season rated #13 in the country. Five weeks earlier the same Michigan team had come to Mackey Arena and won by one, against an expected Purdue margin of 6.9. The road win was the better result by a wide margin, and it is the way the strength rating is meant to work: an opponent is judged by its whole season, so a win over a team that later proved itself is worth more, and a loss to one costs less. One correction to the memory: that Michigan team lost in the Elite Eight. Its runner-up finishes were 1992 and 1993.
The greatest Boilermaker
There is no single answer, and the site refuses to invent one. Instead it measures three things and shows them side by side. The headline lens is a player's share of his team's production, points, rebounds, and shooting efficiency relative to his teammates, multiplied by how good that team proved to be against its schedule. On that lens the best single season since 1952 is Zach Edey's 2024, then Terry Dischinger's 1961 and 1962, Mount's 1969, Joe Barry Carroll's 1979, and Robinson's 1994. The best career is Carroll's, then Edey's, Dischinger's, and Mount's.
Read the adjustment before you trust the order. On raw share alone, Edey's 2024 ranks #15; Dischinger, who carried nearly a third of a middling team's scoring, ranks near the top either way. What moves Edey to first is that his team was the second-best in the country, and the multiplier for that sits at its ceiling. That is a judgment about what greatness means, that carrying a great team counts for more than carrying an ordinary one, and a reader who disagrees can use the unadjusted column, which is printed beside every adjusted number. Robinson's two seasons keep him out of the career list; his 1994 stands sixth among peaks. Players from before 1952, Stretch Murphy and Paul Hoffman among them, have no season lines in the record book and are listed on their own page rather than ranked against players who do.
The third lens, what happened to the team when a player was gone, has a clear leader: Robbie Hummel's 2010. In the five games Purdue played without him after his knee injury, the team performed about 14 points per game worse against expectation than it had with him. It is the largest such gap in the modern era, and it puts a number on a season Purdue fans already remember as the one that got away.
The coaches
Every coach is measured by the same yardsticks, and the author's relation to one of them, Matt Painter is his second cousin, is stated here so that the reader can watch for a thumb on the scale. There isn't one; the numbers are what they are, and they happen to be good.
From 1985 on, the only stretch where Keady and Painter can be compared on identical footing, Painter leads on winning percentage (68.7% to 65.2%), on Big Ten percentage (64.8% to 60.2%), on NCAA tournament wins (24 to 18), on Final Fours (1 to 0), on average team strength, and on accomplishment per season. Keady's 1985-2005 teams won five regular-season Big Ten titles to Painter's five, with Painter adding the program's only two conference tournament titles. Their teams' average rank in the country is the same, about 33th. Lambert's eleven titles in twenty-nine seasons stand apart, in an era the site can measure only coarsely.
The honest part of the coaching story is March. By seed, Purdue's tournament teams have won fewer games than expected under both coaches: Keady's were expected to win about 25 and won 18, Painter's about 26 and won 24. Measured against our own ratings rather than the committee's seeds, both gaps close to under two games. The reading that fits both facts is that Purdue has often been seeded a line or two higher than its strength warranted, and that Painter's teams have come closer to their expectation than Keady's did. Whether that is coaching, roster construction, or the ordinary cruelty of single elimination, forty games cannot say.
Set against every Big Ten coach since 1985, 65 of them with at least two conference seasons, Painter ranks #8 on winning, #10 on the strength of his teams, and #13 on results beyond what those teams' ratings predicted; Keady, for his 1985-2005 seasons, ranks #17, #23, and #14. The names above them are the ones you would expect, Ryan, Knight, Izzo, Matta, and Izzo alone leads the ledger for getting more out of his teams than their ratings said was there. The Big Ten coaches page has all three ledgers and the all-time one, where Lambert's conference winning percentage still ranks #10 of 150 after eighty years.
Where the numbers stop
Three limits are worth carrying with you. The strength rating cannot see anything before 1985 except through Purdue's own conference games, so the Lambert and Mount eras are measured more coarsely than the Keady and Painter eras, and the site labels every such number. The player measures use only the statistics kept in a given season, so the 1950s and 1960s players are judged on points, rebounds, and shooting, and the modern players on much more; a fair comparison across that line is a comparison on the basics. And the accomplishment weights, how much an Elite Eight in 1969 is worth against one in 2019, are a set of judgments written down in a table, not a discovery. Anyone who wants to argue with them can argue with a specific number, which is the most a site like this can offer.
The tables behind every sentence here are one click away on the teams, players, and coaches pages, and every one of them can be downloaded from the data page.