Critical Attack Bonus caps at 100% (Critical Hit Damage) (Updated May 2011)

[Nightfyre]

Terribly, especially since Hi has an innate critrate bonus that invalidates some of that additional critrate on WS. It's a nice atma setup if I'm going for evasion though.

[Atoreis]

Yes but did the results been reanalyzed since the "fix" to make sure the DA/TA canceling were indeed an error from parser ?

Same question for Atoreis, mind posting how YOU tested please ?

Sorry if i sound aggressive on that topic, but proving those hypothesises is playing critically with parser accuracy, sample sizes and the % you aiming to test.
Variances in results obtained made ppl balanced on one hypothesis more than the other and vice versa for very long time...

Conclusion: i would still declare any TA/DA/multiattacks order checks hypothesis as unproven.

I made a parse with turning after every attack round with 3%QA 17%TA 21or22%DA (QA was 3+%, TA was 16.x%, DA was 2% less then expected but I didn't count like 10 rounds in which I killed the mob because i forgot to finish it with my mule) Sample was small 200+ attack rounds but numbers were far far away from DA/TA/QA canceling theory.

[mdkuser]

I made a parse with turning after every attack round with 3%QA 17%TA 21or22%DA (QA was 3+%, TA was 16.x%, DA was 2% less then expected but I didn't count like 10 rounds in which I killed the mob because i forgot to finish it with my mule) Sample was small 200+ attack rounds but numbers were far far away from DA/TA/QA canceling theory.

You are an idiot for two reasons. First you think you can conclude anything with 200 sample size. Secondly you missed byrth remark that kparser did not recognized rounds properly until last week's patch which also means that your "test", which probably used kparser, was also counting incorrectly multiattacks. Last week's tests on alla showed that any previous theory on da/ta/ka was unjustified besides throuh eyeballing ( 200 sample size is eyeballing) of ot was dpne with kparser.

[Deader2818]

Atoreis has pretty much made it clear hes against any form of 3rd party tools (I think even windower to a certain extent) so not sure if he would even use a parser.

[Atoreis]

Atoreis has pretty much made it clear hes against any form of 3rd party tools (I think even windower to a certain extent) so not sure if he would even use a parser.

Very funny... hahahahah... yeah...

Pchan you are an idiot because you assume and you shouldn't. It's not about I'm against using a parse but I wanted to be sure and I was turning after every round and noting on paper which it was QA,TA,DA or single.

[seiji]

I made a parse with turning after every attack round

people cant read, he manually parsed every round

[Deader2818]

Very funny... hahahahah... yeah...

Pchan you are an idiot because you assume and you shouldn't. It's not about I'm against using a parse but I wanted to be sure and I was turning after every round and noting on paper which it was QA,TA,DA or single.

Dude, I wasnt even attacking you, unless im remembering wrong, you have clearly stated before you are against 3rd party programs so I was only stating that you most likely did not use a parser for your test.

[Atoreis]

Dude, I wasnt even attacking you, unless im remembering wrong, you have clearly stated before you are against 3rd party programs so I was only stating that you most likely did not use a parser for your test.

I stated (incorrectly as I was corrected after in case of spellcast) that I am against bots that makes things like casting stun for you or change your gear to resist/pdt/mdt based on mobs action. I'm also against claimbots and fleebots (especially when they are used against other ppl). I'm not against any 3rd party tools that are kinda answer for lol PS2limitation and SE laziness (windower or spellcast for example).

@Pchan I forgot, I don't think i can conclude anything from 200 sample size. I wanted to make 1000 but it was 3 am and they were making something in power station and switched of electricity for like 5 min and since I have no token for my and my friend's mule I couldn't keep it going. I wanted to make it later but then I read on BG this

The results to the contrary were a parser issue. TA > DA has been confirmed by multiple people now and is absolutely the case.

It's all in the Epona's Ring during WS thread on the Alla Monk forums.

and this is why I haven't tried to parse it again.

So yeah stop assuming too much.

[mdkuser]

In your post 68 you are so much confident that your test means anything that we were expecting that you had done anything serious. Turns out you have 200 rounds. Maybe you should shut up before trying to argue against something when in reality you have nothing to back it up. Lolbg

[Masamune]

Sadly were only 200 rounds test, yielding too high error% for each of yur DA/TA/QA expected %...

On other hand where are reanalyzed tests with "fixed" KParser ? i didnot see them at all on the Epona Ring during WS thread @ Allak, still same old results...

[Byrthnoth]

The post I linked was the verification that the "fixed" KParser accurately distinguishes DAs, TAs, and single attacks.

Pchan parsed himself and his mule both with 0% Haste and Signet Staff swinging at the same monster. He had TA and DA, and his mule had none. Given no difference in delay, no JA uses, etc., they should have the same number of attack rounds in the same period of time. Because the mule (control) has no DA or TA, this will be represented purely by its number of swings. Pchan has DA/TA, so through him we determine whether or not the rounds are being interpreted correctly.

The parse shows that it still has some issue with merging rounds (which is unavoidable considering all the analysis is done purely with timestamps), but apart from that it appears to accurately identify TAs and DAs and doesn't split attack rounds anymore.

[Raen Ryong]

In your post 68 you are so much confident that your test means anything that we were expecting that you had done anything serious. Turns out you have 200 rounds. Maybe you should shut up before trying to argue against something when in reality you have nothing to back it up. Lolbg

nice usukane bro

[Foldypaws]

The post I linked was the verification that the "fixed" KParser accurately distinguishes DAs, TAs, and single attacks.

Pchan parsed himself and his mule both with 0% Haste and Signet Staff swinging at the same monster. He had TA and DA, and his mule had none. Given no difference in delay, no JA uses, etc., they should have the same number of attack rounds in the same period of time. Because the mule (control) has no DA or TA, this will be represented purely by its number of swings. Pchan has DA/TA, so through him we determine whether or not the rounds are being interpreted correctly.

The parse shows that it still has some issue with merging rounds (which is unavoidable considering all the analysis is done purely with timestamps), but apart from that it appears to accurately identify TAs and DAs and doesn't split attack rounds anymore.

I was having some issues with it yesterday, AoA on DRG (and RR and SS), was parsing on *3%* TA on myself and 10% on my wyvern. (only ~500 samples though, but 3% was *waaay* out of the confidence interval for 15%.

[Byrthnoth]

How much Haste did you have?

Here is how he calculates it now. I just realized that my earlier link to pchan's post doesn't work because it was subdefault.

I don't know what default base Wyvern delay is, but the bottom line is that your attacks have to be at least >3.5 seconds apart (>~210 Delay) for TA to be picked up accurately by the parser, somewhere close to 300 is probably much safer. He makes the dT histogram and chooses the maximum substantial latency as the time for one round. With a 480 Polearm, Hasso and Haste are probably pushing the limits of what is reasonable. Your Wyvern's parse may be reliable, though. Iirc, our only real verification of the 15% TA from Apoc was Rog using an unknown method (though likely a pre-patched parser).

Example:

Spoiler: show

Attack 1 - Time 00.0
Attack 2 - Time 01.0
Attack 3 - Time 05.5
Attack 4 - Time 06.5
Attack 5 - Time 07.5
Attack 6 - Time 11.0
Attack 7 - Time 16.5
Attack 8 - Time 17.5
Attack 9 - Time 22.0
Attack 10 - Time 27.5

If you take the difference between those:
Attack 2 - Attack 1 = 1.0
Attack 3 - Attack 2 = 4.5
Attack 4 - Attack 3 = 1.0
Attack 5 - Attack 4 = 1.0
Attack 6 - Attack 5 = 3.5
Attack 7 - Attack 6 = 5.5
Attack 8 - Attack 7 = 1.0
Attack 9 - Attack 8 = 4.5
Attack 10 - Attack 9 = 5.5

So in this ideal-yet-short case, he would assume 5.5 seconds is the base attack round. He can use this and the first hit on a monster to determine the start of attack rounds and just add the hits between them.
DA -> TA -> SA -> DA -> SA -> Unknown

[Foldypaws]

25% in gear, Hasso and haste intermittently. 492 lance.

It's not just that the TA rate was low, the DA rate was low too. (also only 3%, though my only DA was brutal.) Only explanation would be it's splitting rounds, but I thought that was fixed now.


I also recall parsing 15% TA on my wyvern in the past, though it was with a pre-patched parser.

I may try to do a more dedicated parse in the future.

[Byrthnoth]

Well, the way the Extra Attacks part of the parser works was changed. It's going to end up being better for some things and worse for others. The current version seems to be quite nice for controlled tests, but the old version possibly will do better in some situations (maybe where you have inconsistent Haste and borderline, for instance).

[Neisan]

Sound like you made some testing on mob's stats ? Can you post (or PM me) your /checks with ALL relevant infos, and i'll post them in Monster Data Mathy Parts (Credits to you of course).

Was posted for training for 2 weeks, credits would have to go to friend using a damage calculator for ranger... I'd have to see if he still has it though since this was at 85 cap and he's been messing around with kinematic's sheet lately since it was a bit more complete and Rng had been shelved. Don't have token back yet (left it in friend's care) so can't login reliably and he's going on a trip right as I get back... I'll do what I can when I can get on but will take awhile before I get around to testing stats again. I didn't start checking stats until monsters were at max level, would have to test again for mobs at min level now as well... they also assumed mobs were mostly war/war and had stats similar to players assuming players could war/war. Several variables need to be eliminated before I can continue on with it.