AF Upgrades

[ChronicCettness]

SKCNM20 you have no hope of landing enfeebles, they're infinitely more evasive than SKCNM10. Each fight was averaging about 6 minutes as well so 1 troubadoured sleep isn't going to do shit when you've got 4 at the very least to kill.

The solution then is clearly GEO!

[Ramzii]

Hard mode honestly didn't seem much more difficult than normal, we cleared Demolition Squad T3 in 4min, T4 in 6 min, and T5 in 25min. Debuffs should probably be fully landable on everything up until hard, acc wasn't really hard pressed at all.

The PLD said he didn't land a single debuff on SCH when we did Demolition Squad Very Hard for the first time, but he said he could have also been packing in an extra ~80 m. acc, I imagine they might be rarely landable with a GEO or a COR giving M. Acc roll, but he suggests trying the RDM ES + new JA pairing as that might be the only way to debuff stuff.

If the M. Eva and Evasion scale proportionally with the difficulty increase, I'd imagine you're going to need to be able to squeeze out every single drop of M. Acc you can find, on normal mode I was capping melee accuracy with acc+28 dex+60, meat, no madrigals, and I only had 50-60% accuracy with acc+150, dex+74, bream sushi, marcato'd blade madrigal (+the 228 skill from oatixur.)

edit:

We considered trying GEO but we're not really sure it's going to outweigh the benefit of having an extra DD since you're already so hard pressed on time. It'd have to make a significant difference to accuracy to merit bringing, I may try and find out tonight.

[ChronicCettness]

Hard mode honestly didn't seem much more difficult than normal, we cleared Demolition Squad T3 in 4min, T4 in 6 min, and T5 in 25min. Debuffs should probably be fully landable on everything up until hard, acc wasn't really hard pressed at all.

The PLD said he didn't land a single debuff on SCH when we did Demolition Squad Very Hard for the first time, but he said he could have also been packing in an extra ~80 m. acc, I imagine they might be rarely landable with a GEO or a COR giving M. Acc roll, but he suggests trying the RDM ES + new JA pairing as that might be the only way to debuff stuff.

If the M. Eva and Evasion scale proportionally with the difficulty increase, I'd imagine you're going to need to be able to squeeze out every single drop of M. Acc you can find, on normal mode I was capping melee accuracy with acc+28 dex+60, meat, no madrigals, and I only had 50-60% accuracy with acc+150, dex+74, bream sushi, marcato'd blade madrigal (+the 228 skill from oatixur.)

edit:

We considered trying GEO but we're not really sure it's going to outweigh the benefit of having an extra DD since you're already so hard pressed on time. It'd have to make a significant difference to accuracy to merit bringing, I may try and find out tonight.

PLD > DD > DD > BRD > RDM > SCH
PLD > DD > DD > GEO > BRD > SCH (let sch and geo enfeeble and sleep at this point)
DD > DD > DD > GEo > Brd > Sch (if enfeebles become reliable screw the tank and sleep em all!)

These are just some ideas... for an average fight vs 4 mobs...
I feel geo's hold their own weight but if you're gonna stack support combo's can't relaly pull DD out else it's not as worth while to stack up just 1 dd :/

[Ramzii]

2nd setup is what we're most likely going to try tonight, but there's no need to sleep if you've got a PLD, they can hold them fine.

EDIT: Figured out the best solution: PLD BRD WHM MNK MNK THF. Feint will more than likely cap accuracy, PLD can hold mobs fine, debuffs are realllllllllllllllllllly not that necessary, losing too much to try and land them.

[ChronicCettness]

Well we did Kindred Spirit's Very Hard and Factory Rejects Very Hard.

Constant Set - Up : Pld , War, War, Cor, Brd, Sch
Kindred Spirit's : Won
Factory Rejects: Wiped twice on the third Doll and Timed out.

Notes about Factory rejects: dolls cast, En-effects from third doll also applied a hate reset on who ever ate the effect. In most cases it was our PLD. Causing our DD's to eat shit fast after.
I'd highly recommend changing set up to : Nin , DD, DD, Geo, Brd, Whm

Cureskin, Arise, stronger bars etc etc. Also /nin would be viable? Just to prevent the hate reset and extra protection vs tp attacks.

Kindred Spirit's I recommend killing BLM > DRK > WAR > SMN. Our pld super tanked them on the stairs. Can use intervals between mobs for recovery/rebuff time. Had a few close calls between the Ice spirit nuking with -ga's and the BLM casting various spells on the DD. No noteable drops that haven't been seen yet. (lvl 109 blue sword?) I'll post screenshots later mayhaps.

Geo could be super useful... so don't discount their utility!

[CDF]

Going back to the topic of crest drop rate, I'm trying to get an overall crest drop rate based on zoning out. It will take a while to get a precise estimate though.

76 Firth Umbrils (good to one-shot with alternating Ebullience Stone V or Fire V, no weather/day bonus required if you have sufficient gear, to avoid monotony of having to kill continuously).

Observed drop rates (with Ionis) (Waning 81% -> 79%)

             Any crest: 24/76 = .3158 (.2139, .4325)
Sacred Kindred's crest:  8/76 = .1052 (.0465, .1970)
           Umbril ooze: 35/76 = .4605 (.3455, .5787)
         Leafkin frond:  5/76 = .0658 (.0217, .1469)
          Dark crystal: 55/76 = .7237 (.6091, .8201)

Suppose the crest drop rate is 25% (conservative assumption). If the distribution of KC/HKC/SKC were 40/40/20 then it would take on average 160 kills to get 8 Sacred Kindred's seals (8/.2/.25 = 160).

Actually it took me 62 kills to get 8 Sacred Kindred's seals (didn't get any SKC for the last 14 kills), which would be a rare event (on the order of 1 in 1000) under the above conditions.

If the distribution of KC/HKC/SKC were uniform then the expected number of kills drops to 8/(1/3)/.25 = 96, which is more plausible given the data.

If the drop rate for crests is 40% and the distribution of KC/HKC/SKC were uniform then the expected value drops to 8/(1/3)/.4 = 60.

Comparing to other people's results, perhaps lockout effectively skews the SKC drop rate downward (if there is a longer timer for SKC vs other crests) in the scenario where people are killing rapidly without zoning. If/when I get more data, this hypothesis can be tested. 93 of 490 crests (source) being SKC (95% CI [.1560, .2274]) suggests either that the SKC percentage is really 20% or there is a longer timer for SKC.

Conclusion: For now I am inclined to believe the crest drop rate distribution is uniform but this needs verification. In any case, zoning seems to help if you don't have TH.

Raw data (png). There are also some timestamps for crest drops to get a sense of whether zoning actually helps (I always zoned between crest drops):

Spoiler: show

[Lyramion]

Thus starts the ara of the Umbril Aery !

[guyonthecouch]

For BST with Pet out, the following looks nice:

Totemic Helm: Pet Acc +20
Totemic Jackcoat: Reward
Totemic Gaiters: Pet Acc/Att +10

+2 legs still best with Stout Servant +1
Monster Gloves +2 with Beast Affinity Aug will always be best with pet out

[pancakesandsex]

Going back to the topic of crest drop rate, I'm trying to get an overall crest drop rate based on zoning out. It will take a while to get a precise estimate though.

76 Firth Umbrils (good to one-shot with alternating Ebullience Stone V or Fire V, no weather/day bonus required if you have sufficient gear, to avoid monotony of having to kill continuously).

Observed drop rates (with Ionis) (Waning 81% -> 79%)

             Any crest: 24/76 = .3158 (.2139, .4325)
Sacred Kindred's crest:  8/76 = .1052 (.0465, .1970)
           Umbril ooze: 35/76 = .4605 (.3455, .5787)
         Leafkin frond:  5/76 = .0658 (.0217, .1469)
          Dark crystal: 55/76 = .7237 (.6091, .8201)

Suppose the crest drop rate is 25% (conservative assumption). If the distribution of KC/HKC/SKC were 40/40/20 then it would take on average 160 kills to get 8 Sacred Kindred's seals (8/.2/.25 = 160).

Actually it took me 62 kills to get 8 Sacred Kindred's seals (didn't get any SKC for the last 14 kills), which would be a rare event (on the order of 1 in 1000) under the above conditions.

If the distribution of KC/HKC/SKC were uniform then the expected number of kills drops to 8/(1/3)/.25 = 96, which is more plausible given the data.

If the drop rate for crests is 40% and the distribution of KC/HKC/SKC were uniform then the expected value drops to 8/(1/3)/.4 = 60.

Comparing to other people's results, perhaps lockout effectively skews the SKC drop rate downward (if there is a longer timer for SKC vs other crests) in the scenario where people are killing rapidly without zoning. If/when I get more data, this hypothesis can be tested. 93 of 490 crests (source) being SKC (95% CI [.1560, .2274]) suggests either that the SKC percentage is really 20% or there is a longer timer for SKC.

Conclusion: For now I am inclined to believe the crest drop rate distribution is uniform but this needs verification. In any case, zoning seems to help if you don't have TH.

Raw data (png). There are also some timestamps for crest drops to get a sense of whether zoning actually helps (I always zoned between crest drops):

Spoiler: show

Could you post your nuking set?

[guyonthecouch]

The mandies in Ceizak drop them pretty regularly when I'm farming on mnk/thf

I farmed over 40 in about ~4 hours of farming

[Sechs]

For BST with Pet out, the following looks nice:
Monster Gloves +2 with Beast Affinity Aug will always be best with pet out

Am I missing something here? I thought the Aug on Monster Gloves +2 applied the moment you used Call Beast and that you didn't have to keep them on for the effect to persist.
Kinda like legs augment and its Haste bonus on SP1

[CDF]

Could you post your nuking set?

http://www.ffxiah.com/item-sets/315265
MAB +64 total on Hagondes armor. I used Fruit Parfait but I don't think I needed it.

(This is not optimized for tier V... more so for tier 1 and helices with dINT < 100 except I don't have Maniacus Sash)

This gets me over 10k with Ebullience

[guyonthecouch]

Am I missing something here? I thought the Aug on Monster Gloves +2 applied the moment you used Call Beast and that you didn't have to keep them on for the effect to persist.
Kinda like legs augment and its Haste bonus on SP1

The source I originally read from said when you un-equip them your Pet falls 5 levels but can be increased by putting them back on. I really, really hope I'm wrong on this lol

[Sechs]

I'm not 100% sure either, which source are you talking of? Might be right altough it would be a disappointing surprise for me.

[NynJa]

Wouldnt dropping its level by 5 drop its HP, and re-raising it would reincrease its hp? Seems pretty easy to test if its bullshit or not.

[lolBST]

It's BS, you can take them off once you call your pet, been tested a long time ago.

[guyonthecouch]

Thanks, I must have read an old/bad source, good to know with confidence now

[Sechs]

Yai! <3

[dasva]

The source I originally read from said when you un-equip them your Pet falls 5 levels but can be increased by putting them back on. I really, really hope I'm wrong on this lol

That's only the weapon

[Ramzii]

Found Legato Dagger in Kindred Spirits. Guess that makes my hypothesis wrong unless SE goofed again