It's hard to get accurate readings with the equipment they use and the relatively small numbers being measured. The PTO dynos are only good for relative comparisons of tractors on _THAT_ dyno. (As Tractor Vet illustrates) A 3hp error is 10% on a 30 horse tractor and 10% error either way is unacceptable from a scientific standpoint.
Weight, weight distribution, engine torque and powertrain momentum all contribute to pulling performance.
At the temple tractor show I always see a crowd gathered around the dyno as they "test" various tractors. I went and looked at it during the lunch break and noticed that with no tractor hooked up at all, the battered looking needle on the gauge was showing 10 hp. How meaningful was any reading on that dyno? Pretty worthless if you ask me.
Attention to detail can make one run better than the other. You can have a 281 with propane head and this and that, but if the valve job is poor or if the rings aren't sealing it won't do you much good. Carefully assembled and tuned stock parts can make a big difference.