JP Ridill Talk. Need translation

[suji]

If you're assuming 33/33/33, the variations in probabilities in your first test (and the variation between your first test and your second test) are just random error. Average swings for both test 1 and 2 would be exactly 2.00000000.

If single, double and triple attacks from a ridill are of equal probabliity, then average attacks per round is 2. By adding a possible DA before ridill proc is tested, you're effectively replacing an attack per round of 2 (the ridill proc) with another attack per round of 2 (th DA proc), which is no difference.

You're right. I didn't quite buy the reasoning since it seemed like a 100% DA proc would do this but not a 10% proc, but given that you're mixing and matching two procs which both yield an average of 2 per round, it's the same. I ran the thing again at 100,000,000 iterations and of course the error averaged out.

My original goal for simulating that was to see if there was a way that made it easy for us to measure whether or not DA procs first or after, which was sort of accomplished by the fact that it showed you should have a noticeable increase in swing count if it's truly procing afterwards, which is why I didn't think it was as stated in my above post. That said, a lot of people seem to think it does in these discussions. Again, by after, I mean DA would replace a Ridill single-proc.

Not sure why I care so much given that I don't even have a Ridill, but this is something to do when bored at work. ;p

[Ambir]

The change in total swing count is still going to be pretty small if the proc is calculated afterwards. The difference between 2.00 and 2.06 would take a lot of swings to confirm. What could be tested is the ratio of single to triple procs which should be similar if DA is calculated first (whether base ridill is 2.00 or 1.92), but significantly skewed if DA procs after.

[suji]

The change in total swing count is still going to be pretty small if the proc is calculated afterwards. The difference between 2.00 and 2.06 would take a lot of swings to confirm. What could be tested is the ratio of single to triple procs which should be similar if DA is calculated first (whether base ridill is 2.00 or 1.92), but significantly skewed if DA procs after.

Yeah, that's the obvious approach but it's extremely tedious to record that data, unless you're using some parser that I'm not aware of. I was trying to look for an answer in swing count that people could passively record without having to pay attention to every time they swung, which doesn't even work all the time if you start glitching from too much haste.

I was thinking it would be much easier if you could merit with someone else who has equal buffs/haste, and measure your swings compared to theirs (including misses) but I think you're right in that the error in doing that test would probably make it too tedious to perform as well.

[Ambir]

The change in total swing count is still going to be pretty small if the proc is calculated afterwards. The difference between 2.00 and 2.06 would take a lot of swings to confirm. What could be tested is the ratio of single to triple procs which should be similar if DA is calculated first (whether base ridill is 2.00 or 1.92), but significantly skewed if DA procs after.

Yeah, that's the obvious approach but it's extremely tedious to record that data, unless you're using some parser that I'm not aware of. I was trying to look for an answer in swing count that people could passively record without having to pay attention to every time they swung, which doesn't even work all the time if you start glitching from too much haste.

I was thinking it would be much easier if you could merit with someone else who has equal buffs/haste, and measure your swings compared to theirs (including misses) but I think you're right in that the error in doing that test would probably make it too tedious to perform as well.

If you can run at both 0% DA and 20% DA the difference should become apparent pretty fast. One of our guys did some work on it recently, just keeping track by hitting 1,2 or 3 on an open wordpad. and looking at how both double and triple attacks (he's a THF too) affected total swing rates.

http://www.qcdn.org/ffxi/showthread.php?t=50849

Over 430 swings with no DA or Triple Attack his ratio was S/D/T = 32/51/17 or 1.85 swings/round.

For WAR vs DRK, the SA/TA rate stayed about the same
for THF vs DRK, the SA/DA rate stayed about the same

but

adding DA gear to WAR the SA/TA rate skewed by a lot.

Is it possible that DA/TA traits proc before Ridill but DA gear procs afterwards?

[Chupunpupun]

Those test does seem to show that DA/TA does help ridill, interesting. DA merit might actually be worth it.

[Icywave]

BTW, while fun and cute, I would'nt take data too seriously.
The DA chart it throws out is just simple math calculation,
everything in that site is also just regular/known infos without going in depth or supported by tests.

[Chupunpupun]

BTW, while fun and cute, I would'nt take data too seriously.
The DA chart it throws out is just simple math calculation,
everything in that site is also just regular/known infos without going in depth or supported by tests.

I was refering to the NA test.

[Retrieving/C.D/Searain]

Those test does seem to show that DA/TA does help ridill, interesting. DA merit might actually be worth it.

All of those tests are based, in my opinion, on an unacceptably small sample size (heck, even 1000+ swings doesn't quite cut it, I've just been too busy to repeat the tests lately and I've had <400 swings sessions whose %'s were all over the place and apart from each other thus convinced me to go at least for 1k+ consecutive rounds) while the results in one of them are completely skewed (the one where the second warrior increased his DA chance via gear).

Instead of trying to guess Ridill multi-hit distribution on paper, making up awkward simulations generated on mathematical premises that, albeit within the realm of reasonable (they'd make all kind of sense in another context, don't get me wrong, I'm not questioning your knowledge), evidently don't apply to "Ridill mechanics" (lol, I think we're steadily and relentlessly approaching the "thinking too hard" line here) cause based on flawed assumptions concerning the sword (the recurring "33-33-33" rumor which was eventually debunked, at least when the sword is wielded by a WAR, TA's and SA's being equally affected, the weapon TA having priority, etc.) that have constituted the epitome of eyballing applied to FFXI for the longest time and, sadly, turned a lot of misconceptions into ZAM-approved popular belief.

The math whiz's should take my data sets (which were scrupulously and unbiasedly collected, thus should be considered reliable, I could've cared less if the results turned out to be the other way around, I would've just adapted) and waste some of their spare time trying to figure out why, from a mathematical point of view, Single Attacks and Triple Attacks %'s aren't proportionately altered by a steady increase of DA chance, also keeping in consideration the fact there could be two distinctly differentiated varieties of "Double Attack" that interact in dissimilar manner with the weapon, that TA's appear to be overwritten by either the JT or the gear-forced DA's and so forth.

My test stemmed from the desire to understand whether DA would've enhanced Ridill performances (even if on a smaller degree compared to the effects it could've had on a regular Axe, Joyeuse docet), or not since I had a Brutal and I wanted to know if equipping it during the TP gain phase would've been worthwhile, however aside from the original purpose I've always been reluctant to draw conclusions and make mathematical speculations on anything else because I've known all along something odd is going on with the RidillTA/SA-Double Attack interactions (and also because we're still lacking a decisive testing session conducted on a DA-neutral job, even though Phraust's stuff at least gives a hint).

We're in the verge of getting our mules to Lv50, one of them is COR, once he'll be 49 I'll go test the sword again (being careful to reproduce the same basic conditions), while under the effects of Fighter's Roll (and I'm willing to temporarily max out my DA merits for testing purpose, I've been sitting at 10/10 and 9999/10k for a while anyway, can't even see chains anymore) that should (hopefully) clear up a few things.

EDIT: bleh, I ended up writing a new chapter of War and Peace.

EDIT#2: one of my friends happen to be an UC Berkeley math teacher and, besides being a friggin awesome player, he's also more of a "FFXI-science" enthusiast than I am, I'll try to get him involved lol, he might find this riddle (how fitting) interesting.

Also: in before ESL, etc.

[Zarath]

Where are these said data, I wouldn't mind spending my free time in school or else where to figure this out hehe.

Additionally, you won't need to merit double attack, the buffs for V and on top of that XI are considerable increase in double attack proc.

[Retrieving/C.D/Searain]

It was on the otherwiki but I'm moving it to the new one, just gimme some time to reword other people's comments (in order to avoid flagrant copypasta'ing) and fix a couple links.

[Icywave]

I haven't read the Ridill tests on BG yet nor my post was directed to you, was just commenting on the JP website ^^; sorry