A quick way to give yourself an idea is to guess at the efficiency, multiply by the power, then run a representative test.
Example:
Say the ESC is 95% efficient on average
Assume we want to run an average of 5kw
5% of 5kw = 250W
250W of heat......
Now go grab one of those gold 30W resistors (you know... the big power resistors in the extruded aluminum housing)
Run 30W through it
See how fast it hits 65C... then how fast it starts to smoke
Compare the thermal mass of the two items
Think about the cooling methods
Think about the temperature rating of the parts on the board
Then think about 250W applied to a similar object - or better yet... put on safety glasses and apply 250W
Not saying this example is directly applicable... just pointing out that there is a reason why things get bigger as powerhandling goes up. Even 1% efficiency loss can mean hundreds of watts to deal with. A tiny thermal mass means that there is no margin - the heat must be removed immediately else power must be retarded or things will get hot and die.
This is what goes through my mind when I look at something and try to gauge how well it will do at a given power level. A lot of us learned this with the airplane ESC's that were designed for a load profile that goes from no-load at 0 RPM to full load at max RPM (with a high wind speed... ) contrary to the bike load profile of full load at 0 RPM and minimum load at speed.
A tiny ESC combined with a hellishly wound motor is highly probable to result in violent ejection of parts from the board.
But - only a test will tell for sure
-methods