That is incorrect. amirali went allin with 0 chips. I looked at the xml file for that hand and before any player action (including posting of small and big blinds) amirali had 0 chips.
<table at=“33” available=“,” bb=“20” … chips=“0,1630,21683,4370,9775,17386” dealer=“5” … nicks=“amirali,Pr1nnyraid,wildbill244,Gambling,StudentBluff2,slickmonte” sb=“10” spectators=“0” states=“p,p,p,p,p,p” … />
Then player actions followed where amirali (seat 0) went allin
<playeraction action=“all in” at=“112” seat=“0” />
And the pot showing 0 chips after amirali auto-posted the small blind
<pot at=“113” pot=“0” seat=“0” total_pot=“0” />
The same thing can be confirmed for the second hand I posted in my original post. In this case, CrazyGirl was not posting blinds and thus went all in later in the hand before the flop. Note the total_pot attribute prior to and after the all in.
<table at=“38” available=“,1,1,” bb=“20” … chips=“14160,0,4642,5100” dealer=“1” … nicks=“slickmonte,CrazyGirl,fmac973,Gambling” sb=“10” spectators=“2” states=“p,p,p,p”… /> <playeraction action=“raise” amount=“40” at=“19551” seat=“0” /> <pot at=“19552” pot=“40” seat=“0” total_pot=“70” /> <playeraction action=“all in” at=“24108” seat=“1” /> <pot at=“24109” pot=“0” seat=“1” total_pot=“70” />
I don’t have anymore information that I can give. So far, I have only seen it happen on royal tables but seeing it 2 times in not nearly a large enough of a sample to deterministically state that the problem is isolated to royal tables; it is too rare of a event. Maybe it’s a specific state the table is in when it occurs. Maybe it’s the data type for chips (floating point vs integer - floating point may be introducing the error).
Sorry I can’t be of any more help. I can only report what I see and what I am able to investigate.