Even so keeping mind that you've got up to 10 chances (depending a bit on how hard you were hit after going down) it's still 50% of rolling a 20, and if we're gonna nerf the bleed system it's just gonna be another thing to promote NOT going out into a dungeon with a party.
First something on that "50% chance". If the rolls would be in fact random, you are right about a theoretical chance of 50%.
Though, it is quite obvious that the engine for random effects like dice rolls is not that random at all.
True - given a large enough number of attempts, the allocation of the results will be nearly equal. But you might have noticed that there is a tendency to a clustering of results.
The pattern seems to imply that you need to make a quite large number of rolls to leave a certain corridor in the results.
It is quite common to roll the same result 5 times or more in a row. Then, the following results can still lie in a corridor again with a repeat of the same results.
Good if you roll 6 criticals in a row, bad if you roll 20 results <10 in a row.
Well, i take it as it is - but please do not tell me that the chance of rolling a result of 20 with 10 attempts is 50%.
Now to the DC of recovering. I see absolutely no reason in a random DC that is sometimes not even beatable at all. The roll to match the DC is random, so why should the DC be random as well?
As for the very high DC: Yes, some characters will most likely recover all the time. So what? They have a strength there, let them have it! I mostly play fighters - i do not complain that i can't cast spells.
The idea of recovery is good, but there are several aspects that make recovery way more unlikely as it is intended to be. Including Monsters that beat the dead horse.