it is maybe close to impossible to extrapolate to 100m for 800 guys, maybe even 1500 guys, but i think you can do it for 5k/10k guys who are endurance monsters & for whom 100/200/400/800...10k is extrapolable as the "speed decay" factor is trivial, as they don't have any speed worth talking about ( you can't do it for 800 guys as guys like rudy with likely 44-high speed woud need to have 1'39-low ability if it was comparable to the 5k/10k guys speed-decay ) !
for a 13'27+ guy ( like i reckon cam is ) a prelim line of fit is
52.2 / 1'50 ->
3'35.8
7'32.6
13'00.1
27'10.1
& most relevantly
24.70
( the 100m extrapolation won't work )
the predictor only says it works for races possible to run at even pace ( a 100 can't as it's 60m of acceleration + 40m of slowing ), but a 200 can be more closely run to even pace ( how mj ran 19.32 off 10.12/9.20 )
the main factor in the 200 is the bend which costs ~ 0.3s
the 24.7 above, represents a "straight" 200, therefore we add 0.3s
that gives an open 200 of
25.0
now, we have to back-track to a 100
usually it's 1/2'ing ratio but for endurance guys, the 100 woud probably be 0.2 - 0.3s slower than the 1/2 value ->
12.7 - 12.8
i woud expect most 13/27+ guys to have this 100