when people say map on the road, reality is its not mapping its logging on the road.. then adjusting map and reload and relog.
mine has been held on the rollers full throttle, high rpms and the rollers win.... car melts into submission in extreme case.
Nothing like these loads are real life tho. Its always accelerating, never sat static at a held rpm in real ife, so why dyno a dynamic load based ecu on a fixed load basis? Its doable and people have this preference. It has its risks in temperature rise and strain that puts on components. Datalogging real life, real road speeds, gears, throttle positions simulates has its advantages, and downsides.. For high powered cars, low gear traction high throttle positions, dont achieve the requested loads due to mechanical traction issues... to tame these you work up the box to "load" more down, but the more power you have the less gears achive the load (more wheelspin) and the end result is using tall gears and the speeds which can occur then, can be high. (same as we drive em for real when conditions allow) - My own car for one is a pain in the neck to map on the road. (too much power not enough grip)
type of ecu vs mapping technicque also a factor.
load based ecu mapping vs 2D map type mapping....
I've run DTA's on the dyno and done exactly the hold at fixe rpm to dial in ign and fueling (non turbo), but run times become so short as you progress up the rev/throttle opening band that you sit waiting for temps to recover more than you do mapping.
I would'nt use this method on a Bosch ME7.x type ecu however given its configuration.
no right or wrongs here... just preferences...