I was in a similar situation in a $30 (R&A) satellite for the $1,000 WSOP.com bracelet event 2 weeks ago.

In my case, there were 43 seats to give away and about 1.7 million chips in play. That told me that an AVERAGE stack when the tournament ended was about 40K. Once I got to about 45K with 70 people left, I started keeping track of how quickly people were being knocked out vs. how quickly I was getting blinded off.

(Note: This relationship is not linear, because the average player gets tighter as we approach the bubble, both because the bubble approaches and because dumber players are likelier to have been eliminated, but it can help to a certain extent to keep track.)

Anyway, on the fly I was using these different bits of data to slowly tighten my range. With 47 people left I folded AK because I was in 19th place still. More anecdotal information I was able to collect was leaving all 5 tables open so I could watch the action and what types of hands were being shown down. If I saw looser play, I knew I was more likely to be successful at folding my way to a seat. And when I saw someone lose 95% of his stack calling an all-in with an A-high flush on a paired board, I knew I was going to make it. (As it turned out so did the player with the flush, but he sure got lucky at the end that there were even worse players out there.)

Anyway, I hope these thought processes help provide a point of view for you to consider. There will always be someone in these large satellites who will do the dirty work so you don’t have to.

Later, I played in a different satellite for the Main Event, in which only 2% of entrants won a seat. Even though there were more than 1,000 people playing, I had to loosen up because only 25 seats were going to be given away. It’s very different from the first example, in which closer to 10% of the entrants ended up getting a seat.