I would want to see the calculations as to why FJ thinks the bug started at line 184 of the "full betting history" (going up). Or some other explanation as to why they believe there is a bug in the game.
If I am reading the spreadsheet correctly, it looks like the OP's "spins" would never have resulted in a loss with a "cash out" of 0.99 or less. In other words, someone could have bet 260+ times to win 1.99x (winning 99% of the bet, and getting the bet back) and won each of those times. As such, I think it is fairly likely there is in fact a bug, however I would want a more detailed explanation from FJ.
The number of bets presented is small, however one scenario could be that the correct formula should be :
SpinBPS * Bet = Win + Bet
It appears the payout formula during when the OP was playing the new game was:
SpinBPS * Bet = Win
If this was the case, it looks like the OP would have won 67 of his 78 bets during the claimed "bug" (if this is the case, the entire betting history for the new game was likely "bugged", and the OP simply had lost bets prior to when FJ claimed the "bug" started).
Other possibilities would be that the SpinBPS was otherwise being calculated incorrectly.
This game is currently up on FJ's website, and is claimed to be provably fair, so FJ presumably knows what the issue was and should be able to show what the outcomes should have been.
I think if the "correct" outcomes from the OP playing the new game would have resulted in the OP having sufficient funds to make the plinko bets up until he won the jackpot, the entire jackpot should be paid out.
If the above is not the case, the floor the OP should receive is what MadZ suggested above. However there are other considerations that would make me believe the OP should receive more, such as FJ essentially "freerolling" their players, and the possibility that the OP may have made different sized rolls had he had a different bankroll size.
I was a bit confused by this as well. The game is similar to Moneypot, where the bet multiplier will continually increase to a certain point, at which it crashes and you receive nothing if you haven't cashed out yet. "SpinBPS" is the crash point of the round, and "BPS" is the user's cashout point, which can be both set automatically or done manually. Whenever BPS < SpinBPS, the user loses their initial wager.
Both SpinBPS and BPS are only included in the second Google doc, so the first one doesn't really give any info.
I don't think the issue here is the system's provable fairness, although I haven't verified it. The bug was that for a certain number of rounds, the user would get paid out their auto-cashout multiplier regardless of whether it was higher than SpinBPS.