Are all the locos you used for the test identical? That is, similar age, exactly the same weight, wheel back to backs, etc? Even though they are all Hornby locos, there is probably small differences in pickup contact tension, weight or number of hours running time etc.
I have four identical Bachmann Davenport 0n30 critters for future bashing into something worthwhile. Only one can handle a speed of less than 8 out of 128, two can handle 12 and the fourth is a rabbit. They all have different light intensities and motor noise. At this point they are all as they came and have been run straight out of the box but part of the bashing upgrade will obviously have to include a serious look at rolling resistance, wiring, weight and decoder settings.
Do the kit built coaches weigh more or less than the Bachmann coaches? Perhaps you also need to look at making all the coaches identical weight as well.