One answer is that if the difference is indistinguishable even with detailed parsing, that is an answer right there: it is neither a big upgrade or downgrade.
If you really want to know what is going on, however, there are two ways to fix the math.
1) Do astonomically long parses so that even the small changes become statistically significant. I won't be doing this, as before I'd undergo this pain I'd be willing to settle for the answer above.
2) Crank up the knobs. Wear 7 pieces of armor that each drop you by 10 AC and add 3 heroic agility. Of course you need to be able to have such armor, but that is always an issue with parsing.