2/2 Clearly that was the reason I was talking about Abellian Groups earlier: trying to pick up math groupies.
Anyone?
Anyone?
#crickets
-
-
@St_Rev A/B test is showing N variants of website to users, collecting stats on purchase rate from each variant then picking best for p >.95Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@St_Rev Today's puzzle: if 2 tests are run @ same time, each w 2 options, naive approach hands user 1 opt 1-a & 2-a, user 2 opt 1-b & 2b -
@St_Rev Thus we get 1.0 correlation between tests, and we're not really testing them independently. So we want to make tests orthogonal by..
End of conversation
New conversation -
-
-
@St_Rev so we rotate test 1 options w each new customer & test 2 options only once test 1 has cycled. But now new problem: if N simul testsThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@St_Rev then HUGELY long cycle times for test 10 or 15. Might take 10yrs before we switch from variant 10-A to 10-B. Thus my work today to..Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@St_Rev to group tests by "flight" where we consider relative primeness of number of options. Which works great for tests w a true prime #..Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@St_Rev ...but it gets tricky when a test has a non-prime number of options. Anyway...back to it.Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
. Banned in Sweden. SubGenius, Zhuangist, white-hat troll. Defrocked mathematician. Brain problems.