WAR, Strengh and Dex 1srt wall to reach

[Maelwys]

Yes and no. You see the test 3 and 4 are the only one you can see an effect via dex as parry is static.

Tests 2 and 5 also have identical parry.

The testing, whilst appreciated, is a little too limited in scope to draw a complete picture of how much Dex contributes.
Because of the major difference in Test #1's results and Test #2's results, we should be aware that low amounts of Dex will give you little or no benefit since it is overshadowed by the testing's margin for error.
That said, there appears to be a floor somewhere around the 215-225 mark where a noticeable benefit starts. And this appears to get noticeably stronger every 15-20 points of Dex.

Once you hit the ~220 Dex floor, assuming that adding ~40 points of Dex will give +2% parry chance is probably a decent (albeit very rough) conservative guideline.

[Coramac]

Yes and no. You see the test 3 and 4 are the only one you can see an effect via dex as parry is static. Now I am not hating on the test and nin's test is greatly appreciated as a time investment. But the data shows a boost boost that was to big to random chance at that dex range so one tier is at 240~246 and if it follows str it is possible that 217 is a tier and maker of the first progression tier as war sits ~36 points away from it but all the teirs discovered (parry/shield reduction, skill speed) are all sliding scales parry str reduct slides by 2 every tier dex might slide by 4 or 6 or maybe not at all. I can't expect anyone to go out there and with a healer and get punched in the face 2-3 thousand times with different sets of gear on. But if you assume its by design tanks must be able to make these teirs with out sacrificing defense and if not that means that neither tank is suppose to parry as a survival tactic.

Test 3 is affected only by parry. Test 4 is affected by Dex and Parry. Dex added 2-3% parry for the latter. His samples aren't enough to have determine the actual rates individually (we won't get parses that long). Collectively, you can nail down some of the parry rates. 24% for tests 1 and 2.

The parsing is actually much better for Dex than it is for parry. I can fit parry to numerous values. I've only been able to fit dex to 1 but it's hardly conclusive as linear programming is more difficult with step functions and the base rate is unknown.

[Sakasa]

Stuff

Stuff

I definantly understand your points and which is why i removed my post here and the others ones as well pointing to the same as I already considered those perspectives. Its to much trouble with the limited numbers because if you wanted to over complicate and not count either test 1or test 2 as a wash when comparing either dex or parry you could even see a pattern that looks like low dex limits parry (please dont think this). Its a can of worms which while important it would still be theory crafting which leads to some misunderstood conclusions. But the point my post was aimed at 240ish would assume that 216ish is the first tier so the gap is similar to the gap you would have to climb from fresh 50 stats as well. But that point while not small looks like semantics but for one willing to try is testable but not worth this explaination.