A better bubble structure than hand-for-hand play? (Long)

[kniper]

We all know the scenario. A tournament bubble approaches, it is in the interest of fair play to prevent giving any benefit to stallers - players who are methodically timing out to hopefully make the money. The tournament grinds to a halt and every table is played hand-for-hand ("H4H"). Both online and in RL, it is both annoying and detrimental to tournament structure: as play grinds to a halt, the time to the next level does not slow.

H4H play is very cumbersome, and for that reason it is hard to implement fairly. Stars I think only implements the H4H play when there is one left until the bubble bursts. Bodog doesn't do it at all if I recall correctly. PokerRoom I remember does it at about three left to the bubble, making it a very long process. The dilemma is that should one make the H4H window very short, you are less effective in stopping stallers (e.g., in Stars there is nothing to stop stallers with 4 to a bubble burst). Make it too long, and you make the play extremely excruciating at the expense of a fair tournament blind structure.

What I propose is a different way to handle the bubble. It is probably easiest to implement online, but shouldn't require any greater logistics in RL than H4H play. I apologize for wasting your time in advance if some site already offers this or it has already been thoroughly discussed. To my knowledge I have never seen it anywhere.

The core of my idea is to institute a "hand count." At some reasonable point before a bubble (the "money" or a pay increase), begin a hand count from zero for each table. Each table will increment their hand count after each hand. When enough players are knocked out such that the bubble has burst, the hand count of the table from which the last player was knocked out becomes the bubble "pivot" (for lack of a better word). Any table that that has a hand count greater than the pivot has automatically burst the bubble. Any table with a hand count less than the pivot has yet to burst the bubble, and players on such tables are still subject to bust out before the bubble. If someone does, a new pivot will be allocated. For this reason, it is necessary to retain and temporarily record the hand count and chip stack (so as to alleviate tournament placing problems when people bust out on the same hand count) of every player that busts from the initiation of the hand count until every table has passed the pivot.

That last paragraph might be a little arcane, so let me clear things up with an example:

Tournament with 400 runners, 10 players/table, 40 pay. We begin our table hand count at a reasonable number before the bubble, say 44. Players bust out from their tables with the following hand counts: 3, 5, 11, and 15. Thus the pivot=15, and we check the 5 tables to see if all the tables have surpassed the pivot point. Say 3 have passed the pivot, but 2 have not. Those 3 tables that have can be consolidated (if possible); they have passed the bubble. Of the remaining two, say one only has a hand count of 8. A short stack busts on that table. Thus, we reset the pivot such that pivot=11, because our bust outs are now 3, 5, 8, and 11. Once the remaining tables pass a hand count of 11, play resumes as normal and their players can be consolidated with the other tables.

If you were spry while reading that last paragraph, you would have noticed a the problem with the hand count method: while a table has not reached the pivot hand count, tables cannot be balanced or consolidated. This limits the window that you can implement the hand count method. For an extreme example, imagine you began a hand count 10 players before the bubble bursts. On one table, one red-hot player busts 9 others by the 5th hand. He then sits by himself, since it would not be fair to move players from other tables unless they had the same hand count, and all the other tables have a hand count > 5. This example is obviously a little over the top, but you can imagine one table with 5 players and another with 9, unable to consolidate unless they have an equal hand count.

There are ways to combat this, such as making tables wait for others to catch up in hand count if table discrepancies get too large. This sort of runs counter to the point of instituting the hand count, so I would advocate keeping the window in which the hand count method is used reasonable, i.e. 3 or so. But it is possible to tinker with the logistics here to come up with a fair way to implement consolidation.

Anyway, sorry for the long read those of you that made it to this point. I would appreciate any feedback as to the functionality of the hand count system or its potential affect on poker bubble strategy; it is entirely possible that I have overlooked a glaring hole in this system. While there is some tricky terrain in implementing this system, I think that as a whole it is much more efficient than H4H play and lends itself to a better flow of the game.