Just about every dyno in the industry displays higher than actual numbers. The history of that goes back a long way (the mid 80's), but at this point high numbers have become an industry standard and machines are (software) calibrated to display a wheel HP number PLUS a certain % (we will call that the "industry standard"... it's very close to what predicted driveline losses should be), and then a correction factor on top of that. I can actually turn the industry standard off on my dyno.... it doesn't matter to me. But if I do, and now the customer thinks his brand new ZX10R makes "only" 150hp stock, and maybe 170ish after a flash/pipe; he will be in denial and just go to another dyno/forum/expert to get the number they think they should have. It's depressing, but it's so wide spread that everyone has to do it to be "on par" with everyone else. Actual, real wheel HP numbers can be calculated easily on my dyno, but no one would believe them after 35+ years of misinformation. At the end of the day, though, it doesn't actually matter what number it says so much as what it says before and after. It's a comparative tool, not an outright measuring stick.

And you are spot on with them being able to be manipulated... that extends to things like tire pressures, engine temperatures, strap technique... there are lots of ways to skew results if you are not careful and consistent. It gets even worse when you start playing with forced induction......
Sadly, the only correction factor I can run on mine I believe is STD, not SAE... that accounts for another 4% or so... and we really don't account for humidity (being in the high desert that's not much of an issue). Temperature is the big bastard and accounts for most of the DA/CF on any given day outside of the static elevation. I always explain all these things to customers, but it can get lost in the sauce with the excitment of being in the room and seeing all the numbers. I will say that this particular dyno is extremely repeatable, and by applying correction factors appropriately the same bike will make within 1% of its power in wildly different atmospheric conditions on runs that are months apart. I keep a pretty big library of comparative runs too, so that bike-to-bike comparisons are valid. I don't like to say "X bike makes Y more power", but rather prefer to say "X bike makes y% more power"... it gives a good real world comparison.