So I have been running reports to try and get an idea of how various 3bet and 4bet sizes perform with different types of hands. One surprising thing that I noticed was that AA usually seemed to perform best when min-3bet. I eventually began to find other strange behavior.
Here is how to reproduce the bug:
Make a report. Select "Preflop 3bet Relative Size" from the "Available Stats" dropbox. In filters, select "Raised Preflop Any 3bet", "Faced All-In Preflop" and "Hands: AA" in any database of reasonable size. You will see that in the report, it lists "3bet < 2x Open" as the most common 3bet size. As far as I can tell, this is happening because it considers the first raiser's 4bet to be the "open size" instead of the original raiser's 2bet. Since the 4bet is obviously larger than the 3bet, it will always consider these spots to be in the "3bet < 2x Open" category even if the 3bet was actually 3x, 4x, or 5x the size of the 2bet.
But it gets weirder... now add "3Bet Size Multiple of the 3Bet Between 2.50 and 4" to the filters we already have. In the report, it will still list almost all (literally all?) of the 3bet sizes as "3Bet < 2x Open" even though this option has been filtered out. I suppose this could mean that the "3Bet Size Multiple of the 3Bet Between 2.50 and 4" filter is working correctly, but the "Preflop 3bet Relative Size" stat in the report is not working correctly.
I believe this problem affects the "Preflop 4bet+ Relative Size" stat as well.
This is a potentially very powerful feature in PT4 that I have been using for many hours, but only now did I notice that this these "relative size" stats being broken has been leading me to false conclusions. This obviously will affect all hands... not just AA. I was just using AA as an easy way to reproduce the bug and share my understanding of what may be going wrong (comparing the 3bet to the first raiser's 4bet rather than his 2bet).
Thanks for any help, and my apologies if I am messing something up myself... but I don't think I am.