[Masamunai]

Hello everyone.

After i tested pDIF caps on lvl1 mobs (pDIF caps test @ BG), i started Raging Rush crit% test, results displayed in the excel sheet at Filefront http://www.filefront...ush%20Test.xls.

For the moment i only did 200ish WSs and only @ 100TP, but when it will get to a point we can confidently know crit @100TP, i'll go to 200TP.
So far the error is still too high but the crit% range makes already the hypothesis "RR crit% @ 100TP=0%" false.

Regarding testing conditions, i did all those WS at exactly 100% TP, i discarded like 20ish WSs i didnot pay attention and went out @ 101TP...
To reach 100TP, i had to work out a way with /sam storetp and bow arrows with exact delay to get exactly 100 as fast as possible (couldnot use meditate in such combo).

If someone want to contribute i'll gladly update this post to get a lower error%.
Also, feel free to point if i did any errors on calculations, i'll fix.

Edit: be careful with calculating normal high and lowest crit'd WSdamage, make sure to use new pDIF caps for 2H weapon written in the excel sheet or @ BG thread in the link above. Might as well attach your results in your post if you not sure.

[Kirschy]

Masamunai, on 03 October 2009 - 10:07 AM, said:

Hello everyone.

After i tested pDIF caps on lvl1 mobs (http://www.bluegartr...2h-weapons.html), i started Raging Rush crit% test, results displayed in the excel sheet as attachment.

For the moment i only did 200ish WSs and only @ 100TP, but when it will get to a point we can confidently know crit @100TP, i'll go to 200TP.
So far the error is still too high but the crit% range makes already the hypothesis "RR crit% @ 100TP=0%" false.

Regarding testing conditions, i did all those WS at exactly 100% TP, i discarded like 20ish WSs i didnot pay attention and went out @ 101TP...
To reach 100TP, i had to work out a way with /sam storetp and bow arrows with exact delay to get exactly 100 as fast as possible (couldnot use meditate in such combo).

If someone want to contribute i'll gladly update this post to get a lesser error%.
Also, feel free to point if i did any errors on calculations, i'll fix.

Edit: be careful with calculating normal high and lowest crit'd WSdamage, make sure to use new pDIF caps for 2H weapon written in the excel sheet or @ BG thread in the link above. Might as well attach your results in your post if you not sure.

This is good info /w well collected data. My only gripe is that there's no testing on crit rate which is possibly not capped?(Although very unlikely!) It does deviate from the expected/aassumed +0% crit rate when using the WS. Interesting results, I'll be happy to see more~

[Masamunai]

Actually i can confirm from the pDIF tests i did already that pretty much all of them were hovering around 22-25% crit%, sample varying from as low as 200ish to 900ish, so can safely say base crit% capped @ 24% (20% base + 4% from merits).

Edit: actually i did even better to answer Kirschy question : i added another sheet with all #crits / #hits from pDIF caps tests, deducing from them the base capped crit rate @ ~25 % with +- 1.1% error.

[Byrthnoth]

Good job!

[Masamunai]

Thank you but don't say good job yet XD : error is still to high to get a definitive value for one, and for 2 still missing 200 & 300 TP...

{Please help me out} {Please assist}

[Byrthnoth]

Sorry, 1/22'ed and left my fat PS2 at my parent's house so I can't even use friends' characters.

The crit rate testing was oscillating between +10% Crit rate during Rampage/Raging Rush at 100TP and +0%, and your results currently indicate around +10-20%. It'd probably be worth going until you can tell what the 100TP bonus is +/- 2.5%. It looks like that'd take about 1200 trials the way you're calculating it.

I'll freely admit that I've always been perplexed about how to calculate a standard deviation of a qualitative assessment (like Crit or not). I think you're assuming a binomial distribution (again, I'm not sure about the implications of having only two categories). I'm pretty sure that you can do something like separate your experiment up into 40 trial blocks and then calculate the standard deviation of a population of means. That way might actually end up making you do fewer trials.

It's fantastic that you're keeping track of your 1%TP returns too. By the time you finish testing for 100%TP you're likely to have enough of them to say something about the additional hits of Raging Rush that could generalize to other multi-hits, which may actually be more interesting than Raging Rush's crit rate itself.

[Masamunai]

updated sheet results.