Re(2): The three stages of Jai-5 jackpots Posted on June 11, 2023 at 01:37:22 PM by Craig G
I completely agree with ... the 3 stages
Well, I don't. That's because with case 3 Tiger states, "(we've not been there yet)". From that it follows that if we are to keep things reality-based, then with this model we have reduced the Jai-5 to just two stages.
But it seems obvious to me that the betting behavior at just over $2,000 will be radically different from when it gets significantly higher, like north of $5,000.
Bear in mind that if there are days with 2 of the games having a dominant player in post 1 or 2, the probs for each could be as good as 1 in 3. Combine that ultra-conservatively with the other 3 games being basically neutral where post 1 or 2 is about a 1 in 6 win chance and we get at least a few probs of approx 1 / (3 x 3 x 6 x 6 x 6) or about 1 in 2,000.
As a point of reference, even our baseline perspective with games where all entries have equal ability, the calculation would be is 1 / (6.13 ^ 5) or about 1 / 8,660 for combo's containing just posts 1 and 2. But when do you ever see a card with no major skill imbalances?
Another point of ref would be a case where there were no heavy faves, but each leg had at least 1 in 5 shot. That still produces combo's that are about 1 in 3,000 hit rate.
What I'm getting at is that the bet is turning into an genuine overlay for a decently skilled hcapper once you get near $5,000. So naturally you should expect to see sharply increased participation when that kicks in.
However, if that is where the competition kicks in, then you have an increased chance of no jackpot owing to multiple winners. So that aspect of game theory muddies the waters.
Bottom line is that stage 2 should be subdivided because it lumps underlay ($2,000-ish) and overlay cases together, and is blind to that distinction.
And one more thing. I think that to analyze the Jai-5 meaningfully, it would be super useful to establish a day-to-day estimate for the combined prob of the top N picks. Very simple and objective barometer. For those computer model guys, that should be child's play.