Make a new spreadsheet.

On sheet 2, generate a lookup table with Ken Pomeroy’s rankings – team in column A, ranking (labeled “Pyth” on KenPom’s data) in column B. At the bottom, add a listing for team “Bye” with ranking 0.

On sheet 1, start at cell A64. In each column, put this data:

A64: name of 1st seed in the upper left bracket as it’s written in KenPom’s data

B64: =VLOOKUP(A64, Sheet2!$A:$B,2,false)

C64: “Bye”

D64: paste from B64

E64: =(B64-B64*D64)/(B64+D64-2*B64*D64)

F64: =RAND()

G64 =IF(F64<=E64,A64,C64)

(A tiny bit of explanation: E64 is the log5 formula – given the “true” winning percentage of two teams, it’s the probability that team 1 will beat team 2. F64 is a random number between 0 and 1.)

Copy cells A64:G64 to cells A1:G127. This will make a mess.

In cells A65:A127, put in the name of each team in bracket order. If it’s a play-in game, put the name of the second team in column C – otherwise leave it as “Bye”.

Change the formulas in cells A1 and C1:

A1: =INDIRECT(“G”&2*ROW(A1))*

C1: =INDIRECT(“G”&2ROW(A1)+1)

Copy cell A1 to A2:A63 and cell C1 to C2:C63.

At this point, cell G1 is your predicted champion of the tournament. Cells G2:G3 are your finalists, cells G4:G7 are your final four, etc. (The cells in column G from row X to row 2X-1 are your final X teams, assuming X is 1, 2, 4, 8, 16, 32, or 64.)

Recalculate the spreadsheet (Ctrl+= in Excel) to get another bracket. Either repeat as necessary or convert the whole thing into a macro/script/program/whatever to get the results you need.

(By the way, these brackets will probably have slightly too many huge upsets – KenPom seems to predict those more often than they occur. Feel free to modify the values in column F to make things more or less predictable.)