It's a sliding scale kind of situation. At one end you have... I don't know... One of your 3rd cornet players missing out a top G because they can't reach it and it's doubled with the whole cornet section and shouldn't have been given to the 3rds anyway; while at the other you have a band playing something completely other than what's expected - a whole different piece, maybe. The former is obviously no kind of problem to anyone, the latter an obvious strangeness.
What on that scale is acceptable? Well, the presentation of a band to an adjudicator is not about individual players, notwithstanding solo spots... It's about the whole ensemble effect. You could probably make an attempt at quantifying what changes were permissible in a particular piece to a particular adjudicator on a particular day for a particular band and MD, but every occurrence is a special case. Despite contesting being an attempt to make sport, this sort of thing is hearteningly much more an art. To pick an unsubtle example, some adjudicators will nod appreciatively at a tastefully and discreetly added tuba pedal - others will drop marks for it. Most often, aural trickery simply goes unnoticed - or perhaps the adjudicator does notice but does not think it worth making anything out of (or if they do, aren't certain of their diagnosis in the absence of visual confirmation). There are many sections in "test-pieces" that are scored in ways that don't come over well with the players available - whether to be deliberately difficult, or because the writer didn't think about it enough, or maybe because it was scored for a specific band whose talents lay in different directions to those of the available players. It is common practice to rescore such sections, minorly or majorly, so that the change is not obvious in the heat of the moment even to the carefully listening ear, and so that a better general ensemble effect is produced.
With reference to this specific example, I don't know the Herbert arrangement in great detail myself, but I see that all of those who are paying closer attention to it here and elsewhere that I've looked find it rather substandard work. One could undoubtedly rearrange the whole suite to more satisfactory effect, but that would be clearly over the line - I mean, one or two adjudicators out there might love it and applaud the intention if executed and played very well, but I would hazard that many more would write comments along the lines of "Is this really the same score as I'm looking at?". It's a plausibility thing - if the adjudicator ever looks at their score and thinks "This isn't what I'm hearing", then you're in trouble. But a surprising amount can be altered to suit without that happening. Another example: St Magnus at the Areas a few years ago; there's a section where the whole back row play quiet and exposed octave slurs Bb to top Bb - not together, but one after the other. Scoring asking for trouble, you might think! We had our annual Area piece workshop on it with a top MD, and we played through that bit - to their credit our back row played it well. Top conductor (also a top adjudicator, note) said something along the lines of "Very good - but we don't need to risk that", and then proceeded to instruct the sop to play the first one, and then hold the top note while all the others simply missed the top note out, only playing the bottom note! Very very much easier - and safer on the contesting stage - and did the adjudicator on the day pick it up? No they didn't.