It should depend a lot on the structure of the tournament, its toughness and the toughness of your table . For instance, if its a winner take all mtt you should take thin edges ( gamble more), if its a more flat structure, u should value your tournent life more. If its a turbo , u should be willing to gamble too, in order to not only acumulate, but survive.
As a rule of thumb, you should not have a rule of thumb in tournament poker :)

I agree, obviously, but I was looking for some rough numbers about the kind of cEV edges you are looking for in different spots and I can then adjust from there for structure, softness, possible icm etc.

For my example hand, it was a SOFT 1200E buyin tournament, table was mostly soft except for 2 players, good structure (avg end of day 1 = 34bb).
What kind of EV (in numbers of bb) would you be looking for in that spot ?

very interesting topic. Ill add ideas: what is the value of your stack if you pick up that pot right here when you 5b shove a2s? Depending on the table, shoving and getting a fold, OR shoving and sucking out, can just provide you great EV situations later.
In the first case, just fear equity. You show that youre ready to go all in, which is extremely strong in tournament, despite the need for survival. You also got the better of an opponent in a metagame perspective, and Id argue that live tournaments with good structures are a lot about navigating the metagame, and one's own image through the day....
If you get lucky and suck out, yes you risked busting, but doubling up and showing that hand will for sure give you a lot of opportunities to get more chips...

All of that is pretty hard to measure. Basically sometimes you want to avoid thin edges like the plague, because the variance is just killing you (when you bust, you miss future EV situations... Youre done. If the field is very fishy and you had bullets in your stack, it rly rly sucks obv). Other times you want to take a lot of thin edges, and potentially even marginal Ev- gambles.
It all depends on your strategy at different times of the tournament, and how the field / your table is playing at that particular moment.....

Yep agree with that, all good comments about psychology and metagame of live mtt and also the future EV of having a huge stack is a factor that is often underrated IMO.

But as I am very mathematically inclined, I'd love to be able to use a program like ICMizer, see the EV of the play, and decide what's the lowest EV I'd want to take the play.

To me it feels that in my hand example (I was already 3rd stack at the table if I folded to his 4b, and the table was not fighting back my opens so far, also we were only playing for less than 2 hours more that day) the benefits of shoving and him folding or shoving and sucking out don't outweigh the costs. So that means that I don't want to take the play if it is 0EV, I want some kind of edge, but I really have no clue where the line is, +0.5bb +1bb, +2bb ?

What do you think about my statement putting the 1bb EV icmizer gives for this one spot, into 100 bb/100 and comparing it with our theoretical bb/100 in this tournament at this point with a 50bb stack (I guesstimate it to be around 20bb/100 FWIW). Can you spot a flaw in this approach ?

Lucas Greenwood made a very good video on this topic. Highly recommend it.
http://www.runitonce.com/poker-training/videos/tournament-life/