Don't forget what bill & mike have both hinted at though... the 'old' dyno is pretty effective at over reading (LCR's did dyno reliabley at 285 there, but 265 anywhere else)
shape wise is what you should be comparing rather than numbers in this instance
But don't get cut up about it, if peak numbers are your thing - that's fine.
What would be interesting is a known to be good dyno to compare & contrast with (sadly we used to use stealth, but that now over reads
)
I'll have to dig out the graph where I replotted the data (I had the raw data) for all the major (at the time)dyno runs
Would be good if you can get that for a range of cars