to further elaborate: i think 6-7 hour is a fairly reasonable length for a tournament that has over 300 runners and a good structure. after all this is the biggest and best tournament UP offers on a weekly basis, so it should go a little longer than your average 300 daily person tournament.
the problem with this structure is that when this tourney was only getting ~90 runners we were still seeing that the tournament was lasting around 6 hours. so how did we add over 200 runners to this tournament yet not see a very big difference in time from start to finish?
easy: the majority of the beginning levels of the tournament are simply useless and people are not busting at a reasonable rate. the tournament is not really "starting" until level 10 or so, when people start to actually battle and play pots, creating bustouts.
but you may ask, shouldn't this still result in a shorter tournament when there are less people (as there are fewer chips in play)? well, it would, however there are many large jumps in levels deep in this tournament. the further into the structure we get, and the more we will experience these massive level jumps...which increases the bust out rate of players deeper in the tournament.
let me try to show what I mean.
with the structure as it stands, we're seeing a lot of players lasting a very long time. once the big jumps in levels begin, we see a very large increase in the rate players are busting...the problem is this rate of increase never slows, as the blind jumps continue throughout the late stages and final table, creating a bustout rate that looks something like this:
(please excuse the fairly crude drawings, it's just to get an idea)
now, if the structure was more ideal...we shoudn't see too many bustouts at the immediate start, but they should start very quickly. the rate players bust should start increasing almost immediately as blinds get larger and antes come into play and by the mid stages we should be seeing players bust fairly quickly...HOWEVER, once we start hitting the bubble and money, bustouts should actually begin to SLOW, as 1 player busting represents a much larger % of the field, therefore we should not be losing them less often. also, we want more play for the final couple of tables and especially at the ft, so we really don't want to be losing players as quickly as we were in the beginning and middle stages.
anyway, it should look something closer to this:
a lot of new pokerrooms (especially in the nj market) seem to think that more starting chips = better structure. this is simply not true, and these graphs represent this very well.