Tournament Strategy Design

One way to model tournaments is to look at the blind levels and have an assumed/estimated rate of chip accumulation. You may accumulate more on different levels based upon your expected chip stack in big blinds, the size of the antes and what stage of the tournament.



Then you can determine you need to double up 3 times throughout the tournament in about 187 hands. You can then determine the probability that you are dealt a given hand X or more times.



Based on this information we need to double up about 3 times given the structure and we have a 64.8% chance of being dealt JJ or better 3 times or more and a 46.2% chance of being dealt QQ+. As you have gone through more hands without doubling up or if you are behind on the rate of chip accumulation you might decide you need to widen your range. you should constantly adjust based upon projected hands left. You can try to double up more or less than this amount and you could easily be dealt 99 and have it be stronger than JJ in certain conditions plus you have postflop to consider, but it's worth thinking about. So let's say you haven't doubled up and have accumulated chips at the expected rate minus a double up and you have 120 hands left projected. Now you can widen your hand selection. Or perhaps you've doubled up and only need to double up 2 more times, you can narrow your hand range or widen it depending on hands left. This is one way to approach risk. Perhaps you want a greater probability of cashing first and so you want a stronger hand and a greater probability of surviving. Or perhaps you want a greater probability of winning and think you can parlay an earlier double up into a greater edge and less risk overall. You can make adjustments. Perhaps there are lots of postflop situations and can get your money in better. Perhaps you aren't likely to get action with your big hand so you need to widen your range or find a way to increase your edge. But this is a good baseline to start with. Slower structures you can be more selective assuming you can still get paid off when you have a hand.

how to calculate Jam or fold ICM

ICM is a valuable calculation in all in or fold scenarios because it can tell you concretely and objectively how much a decision is worth like jamming vs folding.
I have theories about in MTTs how deeper stacks allow players to apply a skill and the opportunity cost of survival to see more hands and find better spots is also greater, but with short stacks when everyone is short stacked near the end of tournaments or near the end of supersattelites or even on the bubble or in sit N Gos ICM is the most relevant because the other stuff doesn't come into play as much.

I would favor an alternative model that requires a lot of guess work for the tournaments that have more playability but ICM is really relevant in super short stacked situations.

Say for instance there are 13 players left in a super satellite that pays 11 places and 11th it pays 180 and 10th or less it pays 250. You have only 3 big blinds or 3000 in chip value with blinds 500/100 and 1000 ante but there are 2 players with fewer chips than you have. You are dealt QQ. It folds to you in the small blind. Folding is better than jamming. How do I know this?
ICM calculation.

There are ICM calculators online that will tell you how much a given chipstack is worth given you also plug in how much your opponents have so they know what percentage of chips you have. In an example I looked at, if you shove and opponent folds 20% of the time you go from 3000 (which is worth $183.05) to 5100 in chips (which is worth $221.71 in ICM) if they fold. But if the big blind calls and you win you go up to 7100 (which is worth $236.20).
So... if you are 75% to win with QQ
(.20*221.71)+(.52*236.20)+(.28*0)=
(44.34)+(122.82)+(0)=167.16
Since folding is worth 183.05, folding is better than jamming. Limping in for half a big blind might be okay provided you examine the very complex decision tree and possible outcomes of each decision and it proves to be more profitable than folding, but that requires a lot of information that you really can't be too confident about.
Folding QQ not only is better than jamming but making it puts $16 in your pocket vs the alternative. Your opponent should NOT be calling so widely, but if he is it hurts both you and your opponent. There is a weird dynamic where if you and your opponents are even in chips you can jam with any two and they might even be correct to fold aces, but as they widen their hand range they hurt themselves and hurt you too to and that continues as they widen their hand range the point where you possibly would be correct to fold aces if they are going to always call you.

You have to consider this as a possibility and look at other calling ranges and the value and weight them by probability of being correct.
If folding QQ with only 3 big blinds is worth that much think of how valuable folding other hands that you might normally want to push on the bubble with deeper stacks. If your opponent folds a lot more often which is possible you can come up with a different equation and when called your odds of winning might decline but you may make enough when opponent folds for it to increase in value and possibly be worth more than folding. Different payout structures may suggest a different strategy as well. You should be aware of the payout structure given X number of opponents and see how it changes and calculate for varying amounts approaching the bubble and on the bubble or have some method to come up with crude approximations because it can put extra money in your pocket.