I know that in Andrew Seidmen's 'Easy Game' he talked about taking some marginally sup-optimal spots if there was a fish at your table sitting deep. But this was in a 100bb cash game setting, where obviously your ability to rebuy is key. I'm not sure it would apply to a tournament, as the value lost from busting out and not being able to play the fish at all any more must outweigh the value gained from just being deeper with the fish. I don't see much application for this at all in tournamaments. 

this initially reminds me of things I here bad players say in live tournaments "i know im behind but if I win iv got a big stack and can probably make the money" (not calling you a bad player as you are thinking about it rather than using it as a justification)

I never gave it too much thought because of this. my initial feeling is that in a situation where you know you have poor equity especially against multiple all-ins where our equity will always be even poorer, I don't think there would ever realistically be enough of an advantage to the stack you would gain.

the reasons for this are that as the situation in which it is likely to occur as you describe would be very early on in the tournament and probably have the worse players still in at this point.

with so long of the comp left stack sizes are going to change and the table dynamics will also, lets think about the outcome scenarios; 

I am thinking of a hand like 910ss with 2 or 3 all-ins in front.

We can fold, in which we keep our 20bb stack- if we are in the first few levels of the comp im assuming we have been coolered once or twice to have this small of a stack (in which case i suggest a tilt/spite-call ha)

We Can call and lose, obviously in which case we are out

We can call and win, in which case we have 80bb ish- this outcome is obviously not a bad one, however this early in the comp are the chips gained worth the risk because of the value our stack size now has? I would lean towards no. of course the stack size will be an advantage over smaller stacks but it isnt by any means going to guarantee a win or even close.
    There is plenty of the early stages left to pick up chips and certainly better spots to stack off, especially if your table is full of bad loose players which the 3/4 way all-in suggests, it would seem in this situation you are likely to have someone willing to stack off when you have the best of it even if that is soon in the first few levels. That being said our 80bb stack now wont have the desired affect of more utility; by this i mean that against these same players fold equity will be no higher because of a bigger stack as they wont fold hands or adjust their play in a way that benefits us.

Obviously the 80bb stack is going to be favourable other to 20bb stack, i just don't think the immediate profitability (at this level) is worth the risk of busting.

I think its a quite common concept used by sng players (18-man mostly) where there's so much value to be gained being the largest of 5 stacks around the bubble, that really marginal +cEV or maybe even slightly -cEV spots in the prebubble fase (7-11 left) are valued more due to future stackutility. 

Not sure if this concept is that applicable in MTTs with the still pretty much non existent moneybubbles there. There are ofcourse other reasons to take -$EV spots in all stages of the MTT. 

I understand what you are saying but I think that talking about negative ICM is unhelpful and will cause confusion. At no stage in an MTT or an SNG is chipEV>$EV. Even HU, winner take all, or in a cash game (same thing) chipEV only equals $EV.

If you can quantify other factors, bounties, your time, chip utility, FPS, elimination of a problem player, whatever - and the benefit of these is greater than the ICM deficit then of course you should act accordingly.

But with 15-20bb I wouldn't be going nuts - plenty play left.

You explained something I've been thinking about for a long time too. Though I don't think it requires a bunch of all ins. In particular, as you pointed out, at key points where we can gain a lot of stack utility that is useful to have for the stage of the tourney I think there is some merit to what you're saying. 

It's difficult to quantity, because if we are on the same page, and I think we are, part of the calculation needs to factor in the skill edge you have. In particular, at around M14-22 (30-50 bigs) there is a more pronounced skill edge awarded to the better players. Between opens, cbets, 3 bets, x/raises and 4 bet decisions for example, the pro is going to make much better +EV decisions than the standards donks in the MTT. These small amounts of continued extra EV compensate for the initial lost EV by making a marginal call at an opportune moment to accumulate a bunch of chips that as you said, offer more utility. If you opt to 'play it safe' you often leave yourself short and have 0 skill edge. A monkey can jam correctly these days.

More specifically, and being results orientated, I had a hand recently where a somewhat aggressive player made a fairly deep jam and I held QKoff. when I used equalab, I think the call was marginal maybe a percent or 2 negative. But rather than fold (arguably correctly and preserve my stack) I knew If I folded I'd be getting short, think it was around M9, with a blind increase imminent, and my skill edge would be soon diminished and my tourney life up to the vengeful poker gods. However by taking the hand (he had AJ I think it was), when I doubled I crippled the only other decent player on the table and gained a significant stack which I was able to leverage effectively in many spots and ultimately won the tourney. 

I've worked extensively with poker snowie and a lot on equilibrium ranges, GTO and in MTT theory. I don't think I've even seen something address this concept you have brought up - If I had more time I would look further into it. I do believe although it can sound a bit fishy, especially to a cash game player, there is more going on than meets the eye in this area you bring up...