dex/crit relation, WS crits, WS gorgets discussion

[Bastilla]

Hello, Im the test Subject in Geno's post. I figured i should say something because i shared the same worries that some of the ppl here. Mainly affraid my hits would drastically go down from changing some key Haste/Att item slots for dex.

The main items changed from my old TPbuild set up were...
Snipers+1x2 +14Acc --> Rajaa/Thunder ring +10DEX
Bomb core +12Att--> Lightning bow +3DEX
Uni. Boots +3%Haste +3STR --> Marine F boots +3DEX
Mithkabobs --> Breem Sushi +6DEX

Before we started the experiment my first worries were Acc since i was taking my snipers for tping i was expecting a wiff fest using meat, but it didnt my Acc% in the parses were (in order of Geno's SS): 88.05%, 85.67, 94.18 Its obvious that with sushi(only reason we used sushi for 3rd test was cuz of the DEX, not my fear of low Acc%) even tho im not wearig snipers that i have way too many Acc, i was a lil shocked to find out my snipers+1 didnt do much however =/

Secondly was my Att, what would the effects of switching my Bomb core for a lightning bow and going from meat to Sushi. First 2 parse were with meat, my DMG%/hit isnt on Geno's SS cuz we wanted to focus on crit rate, but i looked it up myself my avg dmg per hit actually went up from parse 1 to parse 2 by 4 dmg (74.4 to 78.7), but obviously went down from parse 2 to parse 3 by 6 dmg (78.7 to 72.3). Wich brought to my attention that.., well Bomb core didnt seem to do much, but of course meat did.

And last was Haste, well i cant do much to replace that sadly, only thing i can think is sacrificing more att on back slot, forager to commander.

But now what we need to find is the exact number where my crit rate% goes up by to see if i can shade off 3DEX to get the same crit% but put back on my Uni boots for haste or to change from Sushi to meat. I looked for food first on somepage and found this...

Roast Pipira: DEX+3, Mind -1 (OHNOES!) Att+14%(caps:85@607 base) R.Att +14%. 30Mins, Sellable by NPC in Windy Waters for 1kish. HQ of that food fives +4DEX and +14% att no caps.

So more testing is in order! Sorry for wall-o-text =/

EDIT1: Forgot about WS DMG!
For WS DMG the only change i did was keeping Haidates on for WS instead of switching to WAR AF2 pants. I already had considerable +DEX in my previous WS macro from the Hecatomb gear. My WS AVG were...
Parse1: WARAF2pants/Meat: 742 avg on Vorpal Blade
Parse2: B.Haidates/Meat: 768 avg on Vorpal Blade
Parse3: B.Haidates/Sushi: 752 avg on Vorpal Blade

[Xno]

90 DEX is too easy to get , everyone would have noticed a massive crit rate

taru/hume + osode rajas byakko is already 100 DEX

more like 120 DEX on greater colibri camp

Raven crit drop from 29 to 23 while his DEX dropped from 122 to 115

An ideal setup for war would be :

DEX+4 Askar Head ( -1 haste +4 STR Vs Turban )
DEX+3 Lightning bow+1 ( Vs bomb core or Fire Bomblet )
DEX+10 Osode/A hauberk
DEX+5 askar Hand ( -1 haste -5 atk Vs Dusk )
DEX+10 Rajas/Thunder
DEX+15 byakko haidate
DEX+6 Sushi

=> +53 DEX , good for Mithra/Hume/Taru , galka ? maybe ok for elvaann

[RKenshin]

I think I have to agree with Lhexh in the fact that if a WS can't critical at all if TP doesn't modify critical hit-rate, then any kind of +critical% from an item or for DEX is most likely useless for all WS. And I mean this not just for the non-critical WS, but even the WS that have critical hit-rate modified by TP. It just makes the most sense... This would mean your normal critical hit-rate whether buffed by DEX or +critical hit-rate just wouldn't have any effect on weapon skills, making it so the only WS that could possibly critical would be the ones with the TP modifier. And that is exactly what the proof is leading to.

Which, would also mean that DEX and any other critical enhancements are useless for WS. Meaning that using Senju in your main hand would lower damage (for S/F), and that using DEX in any WS (other than Jin since it's modified by DEX) would do nothing more than add accuracy.

But, whether hits other than the first can land critical is hard to test. I'm guessing no, and if I can I think I'll try to test it for sure if I can figure out a way to do it.

Of course, the above is just a theory now until there's more evidence. However, getting true evidence to support this will take an insane amount of testing that I don't think anyone is really up for.

If this ends up being true though it would sort of support what my (and others) results were with regards to stuff like Haidate vs AF2 legs. Ironically it would also prove a lot of people wrong in their belief of loading +DEX for WS. It's hard to say for sure which is really the right answer without more results though.

[Nagamaki]

As was already said, DEX in the 90s is where it seems to cap when fighting these mobs:

Miner Bee 31 - 34
Tavnazian Sheep 33 - 37

I don't think people are overly concerned with dot on massively Too Weak mobs. The tests Bastilla made are a lot more relevant to when it comes to meritting.

The following is pure theory

The different ranges does mean the amount of DEX needed varies less than I would have guessed without any testing. I'll do some tests between 90 and 100 to see where it actually hits 24%, but lets just say its at around 95.

That means you only need around 23 more DEX to cap against VT lv80ish Greater Colibri compared to the TW stuff. Assuming DEX is checked against mob AGI in a straight substraction, that would mean a lv80 mob only has around 23 more AGI than a lv35ish mob.

Elv NIN75/nothing AGI = 55 (wow that's fucking horrible, anyway moving on)
Elv RNG37/nothing AGI = 34

So a gap of 21

I guess its possible mob stats go up somewhat like those of SJless PCs? The exact value could differ, especially considering Elv AGI is pathetic, but the variation seems roughly similar. This is in no way fully certain, but it seems like something that could be used for tests.

[suji]

But, whether hits other than the first can land critical is hard to test. I'm guessing no, and if I can I think I'll try to test it for sure if I can figure out a way to do it.

Of course, the above is just a theory now until there's more evidence. However, getting true evidence to support this will take an insane amount of testing that I don't think anyone is really up for.

Like I was saying above, some testing on it had already begun and it's not really that hard. It's just time consuming. You don't need a super huge sample set to get high confidence with this stuff since critical hit rate is so high on level one mobs.

I'm at work and can't do anything other than waste time posting here right now, but what you'd need to do to test hits other than the first is just record your TP from weaponskills. All TP=1 results would be cases where the first hit misses, so you would be testing hits within the WS. If TP is 5 or whatever it normally is, then you're testing the first-hand hit.

It just sucks since your acc will be high and you will hit mostly but if you kill around 500~700 level 1 mobs, you should miss enough hits to get a decent sample size. Or everyone can just go out and kill 100 mobs in Ronfaure and post their weaponD, STR and attack (just so peers can verify that they are capped).

The test THF did was with DE, a WS where TP modifies accuracy. NIN could do the same thing with Ku and Jin. We could even see if adding DEX increases the crit rate of the WS by measuring the crit rate bonus from a "low dex" setup on normal hits vs a "high dex" setup and then doing a ton of Blade: Jins to look for the same increase.

[RKenshin]

EDIT: In response to the above post, just saw it now.

I'm mostly talking in regards to the 1 TP return hits, I think it's already safe to assume that at least the first hit can't critical.

The fact it's so time consuming is the hard part... I've done a lot of messing around on bunnies and I can't imagine it taking less than a few hours of non-stop testing to get any kind of reasonable amount of information. With a 95% hit-rate on average you'll have to do 1000 WS to get a reasonably accurate test size of about 90 WS, which would take almost all day.

Testing on NIN using Blade: Jin also wouldn't be the best good idea I think, since DEX modifiers Jin damage. A heavy DEX build would naturally do more damage than a low DEX build with Jin just because of modifiers.

@Naga's test.

I'm unsure how you could go about checking the VIT of those mobs, but if someone could we can get an idea of what value their AGI would be around. Using that we could maybe figure out the critical hit equation looks like. It would be a lot of assumptions/guessing, but I think it would still come out to be relatively accurate.

I'm still going with what I was thinking earlier too, and that the tier seems to be passing 2x the mob's AGI.

[mikazuki]

i think mob stats are similar to pc's, except their weapon DMG and HP, offhand i can't remember if they get extra MP, it's been a while since i aspired all the mp off anything >.>;

[Nagamaki]

Double AGI would mean that those lv35ish mobs have around 45-50 AGI, while the lv80 ones would be at around 59 AGI.

That's a pretty small difference, even less than if the cap was reached by something like "have 40 DEX over the mob's AGI".




About non-crit WS being able to crit, my opinion based on experience is a very solid no, they can't. The only way hits in a non-crit based WS can crit is with SA or Assassin TA (1st hit only) and MS (all hits).

I've done direct tests on that subject using Viper Bite on THF and using Shoulder Tackle on MNK, and I've just plain killed enough crap to say that it just doesn't happen.

SATA Dragon Kick/Final Heaven with some boosts and minuets is pretty damn reliable damage and no, it doesn't break the mod based cap (excluding Double Attack, which is clear in TP return). If the 2nd hit could crit, you would see SATAd HTH WS break their calculated damage limit once in awhile in any situation with pdif at/near cap, because the 2nd hit could sometimes crit as well.

Also, I use Final Heaven quite a bit when soloing to keep mob TP down, and while the damage varies a good amount, it also never busts the calculated damage limits without DA, even on mobs where pdif is clearly capped. After the countless hours I've spent throwing FH at mobs while skilling guard/seal farming, I'm pretty confident about this result.

So why does damage on Guillotine/Asuran/DE/whatever vary a lot? The game calculates pdif for each hit separately and adds up the damage, and pdif on regular melee hits is always a range, not an exact value. Get a few high values and you get a high WS, get mostly low ones and you get a low WS, and get a mix and you get an average WS.

[RKenshin]

Actually the AGI difference being that small wouldn't be all too unbelievable. The whole reason I thought that the tier might be 2x a mob's AGI is because I remembered how low most mob's attributes are for their level. As demonstrated by VZX's testing and various other tests mobs and HNMs that are significantly higher/tougher don't have outstanding amounts of defense and VIT like they would seem (or at least to me).

Considering the damage you do to something like an HNM, and the damage you do to an exp mob, it's safe to assume the difference in VIT/etc between them is rather big. Which wouldn't leave much room for an exp mob's stats to be much higher than the 50-70 range or so. Which is also why I estimated the tier to be around 120 originally. As the last tests show, it seems to be reasonably accurate assumption all in all. Nothing is for sure yet of course but it all seems to be coming together and making more sense.

Or, at least that's how I looked at it.

[aurik]

merit mob vit has actually been shown to be fairly reasonable (70s, 80s, etc). I assume agi would be similar.

[RKenshin]

Has it? I actually don't remember the exact numbers but I thought I remember some rather difficult NMs/HNMs having a VIT value of around 80 to 100.

If it is then that really complicates things, because it either means that a mob's attributes vary by a large number or the critical hit-rate equation is way more complicated.

[Keno]

Has it? I actually don't remember the exact numbers but I thought I remember some rather difficult NMs/HNMs having a VIT value of around 80 to 100.

If it is then that really complicates things, because it either means that a mob's attributes vary by a large number or the critical hit-rate equation is way more complicated.

According to VZX:

Ulli VIT = 88 or 89

From his thread over here: http://www.bluegartr.com/forum/viewtopic.php?t=17364

Also has Byakko's VIT at 115.

[Zirk]

It's too bad people hardly do lvl 1 lebros anymore, all out dex/haste build would tear that shit up

Those were nice tests Geno/Bastilla, always nice to see that the tests on TW mobs correlate to merit situation. I'm going have to agree though that building a dex gear set, even one that barely breaks the crit tier, will still not be as good as the normal attack/acc/haste in 95% of the situations. I think only time I could see it working in a merit situation is on mithra at a camp with mobs that seem to/somehow were proven to have lower AGI than other exp mobs.

[RKenshin]

@Keno

Yeah I read, I was partially referring to that. I remember there were other tests calculating VIT as well though, and that's the test I'm having trouble remembering. Trying to search for it now but nothing, and actually my laptop battery is about to die so I'm probably going to have to call it quits for now.

[aurik]

basically, merit mobs agi is most likely somewhere between 60-80.

[Nagamaki]

There's a pretty simple way to check if external (ie not the WS's TP modifier) stats affect the crit rate of a crit-by-TP WS, assuming one big thing: does +crit gear/stats affect ranged crit rate in any way? I actually have no idea whether it does.

If it does though... Sniper Shot/Dulling Arrow or Heavy Shot/Arching Arrow (crit WS), point blank, lv0 Wild Rabbits.

I can definitely tell you that the crit rate of Sniper Shot, used asap at 100% (so within 20 of 100%) is really, really crappy. I used that to unlock my Coffinmaker, mostly solo, and I remember crits being extremely few and far between.

So if crit modifiers apply to ranged, that would be a very simple test. If it doesn't, then the sampling size will have to be a lot bigger to test with melee WS.

[Keno]

It's too bad people hardly do lvl 1 lebros anymore, all out dex/haste build would tear that shit up

Those were nice tests Geno/Bastilla, always nice to see that the tests on TW mobs correlate to merit situation. I'm going have to agree though that building a dex gear set, even one that barely breaks the crit tier, will still not be as good as the normal attack/acc/haste in 95% of the situations. I think only time I could see it working in a merit situation is on mithra at a camp with mobs that seem to/somehow were proven to have lower AGI than other exp mobs.

It really depends on what the tier is on merit mobs. If DEX is around 100, it's possible, 110, less likely, 120, even less. But the only real answers will come from solid tests, not speculation.

And for NINs it would be much easier, because they can use U/U. This has already been shown to parse equally if not higher than S/F without considering breaking a DEX tier, so the extra +7 from that is a given.

[Zilong]

This is probably the only instance in my FFXI career where I'd wish that I was a catfag ; ;

[aurik]

U/U. This has already been shown to parse equally if not higher than S/F

don't go bringing that into this discussion

[geno]

Geno,

While the parse results are interesting and show a nice difference in the crit rate with the different builds/food, is that really the number that we should be caring about? In all of those builds, Bastilla ended up doing ~40% of the total damage. His crit rate may have doubled but his damage output was substantially lower, ultimately resulting in the same relative amount of damage. So what does the DEX really matter if it ultimately doesn't improve your performance significantly over a more typical build?

Actually yes, crit rate is the only number you should be concerned with in the parses. There were many other different factors that went into DMG% (Deaths, Lazy BRDs, spending merits, refreshing sanction, and 2 hours (obviously not Bast, but Kat and the BRDs both 2 houred.) Not to mention I thought I made a point to not take the DMG% to much to heart because we were not out to maximize damage potential - just to see if there was a crit teir and at what relative point it was at. edit: ONE MORE THING about dmg. When bast switched gearsets, Kat also switched to never using seigan, using hasso only from that point on. I just learned this, so I just posted it.

To recap, as of now we know at some point between +40 and +49 DEX, Bast made massive gains on critical hit rate. Assuming the actual point this happened is actually 49 and we just got lucky, than yes maybe this build isn't the best viable option (although I guess even thats up to debate if you want to contend that if all melees have the same setup, you could use 2xMinu 2xMarch). But this is where it gets interesting. If the crit tier is 46, than bast can switch to a stronger food. Here let me make a chart of some of the plausible scenarios.

http://img.photobucket.com/albums/v6...L/DEXchart.jpg
(edit: the stats on this do not reflect the bonuses from food)

I should note that the meat I am referring to this chart is any one of a number of types that give +3DEX (+4DEX HQ) and varying ammounts of attack (most of which exceed the attack of a Mithkabob). Some of these foods even give the +4DEX on NQ food, so if those turn out to be viable, than where I put HQ food on the chart, should just be NQ.
Here are some examples:
Blackened Newt (NQ=4DEX HQ=4DEX, Attack+60)
Roast Papira (NQ=3DEX HQ=4DEX, Attack+85)
Shark Fin Soup (NQ=4DEX HQ=4DEX, Attack +70)
Pickled Herring (NQ=3DEX HQ=4DEX, Attack +12%)

Thats all I could find for +3 and +4DEX attack foods. Two of those are sold from an NPC, one is a 3 hour food, and the other is only crafted. I'm not going to get into the viability of each of these as thats out of my scope as a lowly bard to judge what food people want to feed colibiri for a couple hours, just showing that it is possible to keep high DEX with an attack build.

So it is definitely possible to get some high DEX while maintaining a haste build, and even using meat. One thing I didn't even mention on my chart though was using sushi still, but switching in Unicorn leggings, thus wearing full haste+% gear and still having +46DEX. And hell, if you REALLY want to go all out with your gear and leave out some DEX, you could always pump up your DEX merits!
Anyways, I just think its too early to conclude that this discovery doesn't serve any practical purposes in EXP.