On the road it's about strength, power and a good set of lungs, weight only has a smaller part to play, not forgetting bike handling on descents, not one of my strongpoints.
I would say that weight does play a big part, but only on long steep climbs. Just straight power to weight calculations can show how you can go, say, 10%-20% faster by losing excess fat. There is a reason why great climbers are generally scrawny!
Last summer I was fitter than I had been the year before, but had put weight on following time in hospital and found that although I was beating all my strava times on the flat and downhill, I was getting nowhere near on the climbs.
Then last winter I dropped my weight from 82kg to 73kg. Taking into account the weight of the bike and kit, that's about a 10% improvement in power to weight, and it really showed when I got back out onto my favourite 1.3 mile 8% climb.
Only true on long climbs over about 6%, much below that and it is only really power and aerodynamics that count.
The Bkool sim seems to handle weight differences quite accurately on climbs, but on descents I think it is way out.
In reality, on a descent, drag and bike handling (not weight) are likely to determine speed. Weight would be a factor in initial acceleration, but as soon as the speed got up air resistance would be much more important. A heavy person would tend to be larger and therefore have more drag, and to go faster requires exponential increases in force and therefore small differences in mass would make even smaller differences in terminal velocity.
Very quickly riders would reach top speeds mainly determined by how much they can tuck in (and smaller riders would have a benefit here) and how brave they are on the corners (no corners on bkool). The force of gravity on a larger mass would have a much smaller impact than when climbing steeply and slowly.
Cheers,
Geoff