I am looking for a simple script that puts every player all in every hand until the end of the tournament.
Example:
Input:
Player 1: 1000
Player 2: 700
Player 3: 755
Player 4: 612
Number of simulations: 10000
Prize 1st place: 100$
Prize 2nd place: 50$
Prize 3rd place: 25$
Output:
Player 1 finished 1st 29%, 2nd 60%, 3rd 5%, 4th 6%. EV=60.25$
Player 2 finished 1st 14%, 2nd 50%, 3rd 25%, 4th 11%. EV=45.25$
// same for every player.

A simulation means all the hands needed until somebody wins.
Ignore split pots.
Every player has an equal chance of winning a hand (you don't need to actually generate hands and boards, just use probabilities).
Note that in the above example there is the main pot and 2 sidepots and this is very important in case player 1 doesn't win the first hand.
If more players are eliminated simultaneously the one with more chips gets the higher place. In the example if players 1 wins the first hand, player 3 gets 2nd place (50$) and players 2 3rd place (25$).
If 2 players have the same number of chips and are getting eliminated simultaneously they split equally the sum of the 2 prizes and receive the higher place.
These are all normal tournament rules, but I wanted to clarify in case some of you don't know them.

Number of players ( at input) can be between 3 and 8.
Places paid between 2 and 8.
Visuals are not important. It can be just some text like in the example.

If there is already such a program, please let me know, I didn't find anything ( I'll give you a tip).
If you decide to work on this, post below so anybody else will not waste time.

Assuming the probabilities are ICM based, wouldn't an ICM calculator do the job?

The probabilities have nothing to do with ICM.
In the example I gave every player has 25% chance of winning the first hand. If player 4 wins, every other player has 33% chance of winning the sidepot and so on.
An ICM calculator would work only if they allow more than 3 players all in (none of them do) and 100% push and calling ranges for every player until the end (none allow locking future game simulation strategy).

Lets take an extreme example where player 4 has 1 chip instead of 612. If he wins the main pot in the first hand and player 1 wins the sidepot, the guy with 1 chip already has a guaranteed 2nd place (this happens 0.25*0.33= 8.2%). His 1 chip is way more valuable than any ICM algorithm would predict because the other guys are forced all in.

Hmm I actually might make this later on, not sure it's really worth the time but it is kind of an interesting idea I must admit.

Interesting problem, I gave it a first try. Neglecting split pot outcomes simplifies the calculations. Here are my results

For 100000 iterations

On average the initial stacksizes:
{'Player1': 1000, 'Player2': 700, 'Player3': 755, 'Player4': 612}
Average finish:
{'Player1': 1.82022, 'Player2': 2.65053, 'Player3': 2.229, 'Player4': 3.30025}
Average cash:
{'Player1': 62.66225, 'Player2': 39.45625, 'Player3': 50.3905, 'Player4': 22.491}

For 1000000 iterations:

On average the initial stacksizes:
{'Player1': 1000, 'Player2': 700, 'Player3': 755, 'Player4': 612}
Average finish:
{'Player1': 1.817661, 'Player2': 2.652605, 'Player3': 2.227348, 'Player4': 3.302386}
Average cash:
{'Player1': 62.7124, 'Player2': 39.410225, 'Player3': 50.454725, 'Player4': 22.42265}


Don't have time atm to rewrite average finishes in your desired format.

Wow, nice work!
Average cash (EV) is the most important aspect for me.
I wanted to send you a pm but I guess I can't because your account is new.
Can you add andrei.avramia on skype?