Tuesday, September 28, 2010

Weekend update 9

Homam Gambieras:
No screenshot with me, but there's no need, is there? Did Omega with 9 people. New toys like Saboteur and level 85 make it even easier than it was before. I'm confident we can take it out with 6, but I guess there's no harm in being cautious. My magic accuracy while Chainspell nuking the last phase was pretty good. I think once my elemental skill is capped I should be able to out-damage the BLM's in the final stretch. I'm not sure that I don't already, because I didn't parse.

Last saturday it was my intention to farm mega Gold Pyxis for key items in Abyssea Konschat highlands. However, there were two groups at bugards and I didn't know any other mobs that dropped all lights that we could exp on. Since I was kind of stressing with building the alliance in the first place, I had no energy left to look for an alternative and we just defaulted to Shadow Lizards. To my surprise we got a single mega Gold Pyxis despite never having had more than 6 Amber lights built up. It contained the Fistule key item which I suppose we accomplished PART of what we set out to do. In the end it was just a simple EXP party, which isn't a bad thing. I got my DRK to 85. Still have scythe and magic skills to cap and merits to sort out, but I'm getting there. Monday I got this in a Misareaux Colibri EXP group:

Didn't know Dark Magic skill was a possible augment on these. During the rest of the session we got and acc+8 version, but no other ones with dark magic on them. So it's probably rare. The DPS on the scythe isn't horrible. It's higher than Perdu Sickle with latent effect triggered 11.25 vs 11.14. The base damage is low, but so is the delay. While I lack any better scythe options I'm going to be using this. The accuracy is nice until I cap and get merits and the dark skill is a cute little boost. I really want something I can tank with that will enhance the effectiveness of Absorbs, Drains and Aspirs without having to swap out for Pluto's Staff. Until I've figured out what I want, I'm going to use this. It's probably going to end up being one of the elemental Magian scythe paths. The nice thing about the 480 delay on the Zaghnal is that a 6-hit without SAM sub is so far out of reach that I don't have to worry about it and a 7-hit comes rather easily.

The session where I got that scythe also netted me some nice EXP on my THF that jumped from halfway into 75 to 80. I learned Aeolian Edge at 290 dagger skill. No use in trying it out on Colibri, but this morning I did this on Red mage:

That 800 damage was done at 300 TP and at 104 TP I managed 570 damage. Around 100 TP it's much stronger than Cyclone which did 249 with the same gear. Around 200 TP Aeolian did 705 damage which isn't much less than at the full 300. It's better than Cyclone all around, but it gets less of a boost from extra TP than Cyclone does. Martial Knife would be a good option if you were trying to max Aeolian on a regular basis. While NOWHERE near the damage Fell Cleave is able to do, it's still unscaled AOE damage that you can use to mow down large groups of mobs. If you recall my PDT tanking test on mandies in Boyahda Tree, this could even be done solo without support if you find a target worth farming in large enough numbers. Ok fine, I'll admit I can't think of a good application yet, but it's not a worthless WS for RDM's. For a THF... it's pretty crappy.

Monday, September 27, 2010

Popular ways to fight NM's

I remember seeing a decision-tree once that detailed different ways you could fight NM's in FFXI depending on what properties they posessed. All the way at the bottom of the tree, given a certain sequence of answers to the questions posed in the chart, was the ominous conclusion that "You're fighting AV, run!". It was funny, but also a useful reminder of what parameters you need to take into consideration when devising a strategy.

I asked myself the same type of question: In which situation will you use what strategy to fight an NM? There are several types of fights when I consider all the info I've read over the years. I believe the following list covers most (if not all) of these.
  1. Zergs
  2. Unengaged tank
  3. Kited
  4. Straight tank
  5. Pet burn
Named after the Starcraft race popular for an early mass-numbers attack called the "Zergling rush", the idea is simple: Kill it before it kills you. That means do as much damage as you can as fast as you can before the NM is able to take you out. The typical way is to have a full alliance of melee damage dealers and some support, rotate bards/corsairs/summoners to buff them with offensive and defensive status effects and engage simultaneously while weakening the target mob (Feint, Angon, Dia etc.) and trying to immobilize it with stuns if possible.

So when is this a good idea? As with any other strategy: whenever it works. Still, there's more things to consider when making a strategic decision. Sometimes you want to be optimal, other times you need to fight multiple targets and you want to be flexible. The positives things about a zerg that I can think of are:
  • It's relatively simple to execute
  • The fights are quick
  • It's fun (for the DD's anyway)
Down sides:
  • It's all or nothing (it's usually hard to recover from a failure)
  • Not always repeatable (if 2-hour abilities are needed)
  • High numbers requirement
Obviously it's only attractive when you "know" you can pull it off. However, what if you don't know if it can be done? What if there's little to no info about an NM you're about to fight or if nobody's zerged it before? Ask yourself:
  • Is Mob_HP / #Players * Average_DPS less than Maximum_fight_time? If you really want to know if it's doable to the second, you need to have an idea of how much HP an NM has and how much damage you can do to it per second. It'd require some research and estimates, I reckon. Parses help.
  • Can it wipe the alliance? With AoE magic and TP abilities, the typical melee is usually in danger of dying to an NM when left unbuffed. Can you prevent the deaths for the duration of the fight? Defensive options include things like: Stun-locking the NM, Chainspell Curaga, Benediction, Perfect Defense and the new Earthern Armor and PDT bard song.
Unengaged tank
You could say this is the polar opposite of a zerg. The idea is to have one or more tanks building hate without feeding a mob TP through melee and using high damage single attacks to get an NM down slowly. This limits the number of offensive abilities a mob will be able to use during the fight. This defensive type of fight is particularly useful when you're not in a hurry and could risk healers running out of MP or stealing hate if you damaged it any faster. Limiting TP to a mob can be achieved by:
  • Not meleeing
  • Using only high-damage low-TP attacks (Nukes, SATA+Cannonball/WS, Rangers)
  • Absorb-TP, Feather Tickle (?)
  • Not getting hit (Utsusemi, Evasion tanking, bouncing hate, Slow/Elegy/Paralyze)
With Absorb-TP in the mix you can not just limit TP feed, but also remove TP. This leads to some extreme situations where a mob without strong Regain will not use a single TP ability. This is the main reason I think DRK is a strong tank and why I first started leveling it. When will this strategy work out?
  • When enmity can be maintained by tanks. If an Enmity reset can occur, then it's a question of having enough tanks to keep at least one of them above the mages on the hate-list. If the enmity reset is a TP ability and you can prevent it, then you shouldn't need more than two.
  • When a mob doesn't have a strong regain. Low regain is manageable most of the time, especially if TP reduction is being used. However, as with the last phase of the Proto-Omega fight, a quick kill may be necessary once the mob reaches low HP. As far as I know, all mobs start using their TP as soon as they reach 100% once their HP falls below ~25%.
  • When a mob is dangerous regardless of TP. Magic casting mobs that can't be stunned and those that use abilities regardless of TP (HP% or post draw-in) are going to give you trouble if fought this way.
It's a safe way of fighting if you have the proper tank classes. I praise DRK in being ideal for this, but NIN's, PLD's, BLU's, WHM's and even RDM's (still) can achieve this using Cure cheats and other hate building tools. With some help from Trick Attack almost any class with Utsusemi can pull this off.

This includes normal kites and pinning. The idea is to limit the amount of damage a mob is able to do by outrunning it. This limits its damage output (barring mistakes) to:
  • Magic spells
  • Long range TP abilities
  • Moments where it changes targets
Some popular kite fights used to be Kirin and Vrrtra. They were considered too dangerous to straight tank. After it was shown that they were zergable, that became the preferred method of dealing with them. However, mobs like Meteor casting Behemoths have some anti-zerg TP abilities and a high resistance/imunity to Stun. These are still kited for the most part. It's all a matter of being able to outrun the enemy. Which depends on things like how fast they move, Gravity and Bind landing, having movement speed enhancing gear, whether the mob can hinder your movement and whether or not the mob casts magic spells. Spellcasting mobs are generally easy to kite. Still, even when it can be outrun, certain things can put an end to your kited efforts:
  • Draw-in
  • Unfavourable terrain/environment
While kiting you can deal damage through ranged attacks, magic spells, pets and avatars. If anyone is able to outrun the mob, then you don't need any specialized jobs and could just bounce hate around the damage dealers.

Straight tank
This is almost a catch-all category. It's, from my experience, the most used "strategy" in FFXI. It's a simple matter of: toss people on it that can keep hate of mages and keep them alive, a.k.a. put some tanks on it and deal with damage and debuffs as they come along. It's different from a zerg, because you're not intending to make the fight as short as possible. It's not like an unengaged tank, because you're still feeding it a reasonable amount of TP by hitting it. It's obviously not like kiting. It's a middle of the road strategy where success isn't based on some unique advantage you're exploiting, but on a simple heavy-weight slugfest. It's a compromise between safety and a fast kill and for that reason it's effective on the vast majority of NM's in the game.

I don't see a strategic difference between using a Paladin and a DD tank, so I lump it all together. The traditional tank may have more survivability and hate tools that don't depend as much on level difference, but I don't consider the difference in means to maintain hate as something very fundamental. Your defense lies primarily in not having too many people feeding TP and taking damage as well as keeping your healers' MP pool sufficiently full.

When using DD tanks, I assume the target is easily meleed. In that case you don't necessarily need additional sources of damage. However, if you're using Paladins or another class that doesn't deal lot of damage, the fight could drag out. Hitting the enmity cap could make things a bit sticky and a healer could get pummeled. It shouldn't make the fight impossible, but adding BLM's, SMN's or RNG's would help speed it up.

Pet burn
Not very popular due to the jobs required to pull it off, pet burning can be a very effective strategy. When death of a tank is part of the plan, a mob's offense suddenly becomes less threatning. The greatest risks involved in pet burning lies in the speed at which pets can be replenished and the balance of hate between pets and masters. The replenishment can be compensated by enhancing the pets' durability (through gear) adding more players and mixing up BST's (timer based) and SMN's (MP based). The balance of hate can be compensated by layering the pets so that they don't all die at once. The short periods where a mob doesn't have the attention of a pet form the only real risk. Even so, when weakened Beastmasters can employ the same strategy as their pets are not affected by their master's weakness.

What's best?
It depends on the mob, but also on what's available to you. A fact that is frequently overlooked by people fortunate enough to have a Linkshell with a huge job selection. I think if you CAN zerg, you should zerg. It's fast, easy and a fun pissing contest for the DD's. If you can't, you should just straight tank it. It's still reasonably fast and it's exciting because there's still danger. If your mob is too dangerous to go head-to-head with and additional defensive measures can't compensate, then my third option is to kite it. Kiting is easy when it works and a very strong tactic. It makes certain solos possible that wouldn't be possible if the mob couldn't be outrun. SE recognises this as is evidenced by a lot of new mobs posessing extremely fast movement speed. If you can't kite it then a very safe unengaged fight is an option. This order is just my personal preference based on safety, speed and fun.

Tuesday, September 21, 2010

Abyssea: Scars

Dual meaning! Last night we tried to get the Doffled Poroggo hat key-item so we could pop Misareaux's zone boss (quest boss). That thing is highly sought-after. the first time I pulled it, it didn't even aggro or come after me, the second time I ran into some big river rocks and died in a hasty pull trying to beat someone else going after it. Still, it wasn't a bad run.

Black Rabbit
Npfundlwa dropped Fortune Abyssite that increases the rate at which Pyxis appear. It's nothing conclusive but based on the rest of the run it seemed powerful. Seal drops where "meh". Though it did help get Magnumm his WHM pants along with another WHM seal drop later on. Rabbit's dust cloud TP ability does very little damage and it only uses that one. It's probably soloable.

Some Bad Pot
Tuskertrap went badly. I popped when I saw there was room, but didn't see someone engage one of the boartraps trying to clear the way for a pull. So, in essence it only left 1 DD tank. It was clearly not enough as it attacks fast and for decently high numbers. I was unweakened but again left to tank solo, after another DD got up they were trying to solo tank again. So we couldn't recover after a bad start and lost the pop. Some people said we needed a real tank, but honestly I think 3 DD/NIN tanks is enough as long as they're all up at the same time bouncing hate. We had a bad start, and that's why it seemed hard.

Minax Bugard has either weird hate or the BLM's were just blasting too much. If there's weird hate at all and it's a hate reset ability then it's probably Heavy Bellow. If it's not, then it's just fast decay or nothing at all. Either way, he's very non-threatening.

I spit on you
Gukumatz didn't seem to cause as much of a problem with hate reset as I expected. 3 DD tanks were able to keep it off mages. I suppose it didn't live long enough to really spam the eyes move that resets hate.

Sirrush spams snowball which means no stun move, no petrification and nothing dangerous at all as it doesn't do a lot of damage. So, again, tier1 NM's prove to be simple.

Nehebkau uses a breath move that stuns for a long time and it gets Shock Spikes, but because we have 3 people bouncing hate, most of its hits still hit shadows. This one put out a SAM legs seal which put me at 3/8. We raced to Heqet (frog) after this fight only to lose claim due to a crappy puller (me).

I don't believe we caught any of the mobs' weaknesses. I asked Tipira to cast all his spells to try and find them, and I'm not sure it worked on any. However, I'm inclined to believe we did get a weakness on Gukumatz because it dropped more than 1 item, which was the tendency the rest of the mobs had that day.

Heqet RNG solo?
I found out that if you do a ranged attack to Heqet from maximum range, it doesn't aggro. The name goes red then yellow after a couple of seconds. If it doesn't move in between shots due to the deaggro cooldown, a RNG could probably make this the easiest solo ever. [Edit: this doesn't work, obviously. It just takes two actions to aggro it. Amun has the same problem.]

Coming Soon
Unable to get zone boss pop set is disappointing, but I believe it should be easier to do this in Attohwa Chasm. Misareaux is probably the most accessible and most popular Scars of Abyssea zone. I hope Attohwa will have little to no competition for the Chigoe NM, which is the only free-spawn NM needed for the quest boss trigger set. We'll also try and get Sojourn Abyssite from the treant NM. Loading up on Abyssites will help us as a group so trying to side-track from gear-dropping NM's for these is worth the time.

Monday, September 13, 2010

The world after "AF3": Samurai

Version update came around 5 days ago and introduced the Empyrean Armor which everyone is referring to as AF3 (as that is what someone from the dev team called it in an interview). The moment DAT miners found the stats on these people's minds have been continually blown. I'm trying to make sense of the implications as it's now pretty clear how attainable these new pieces of armor are.

Like many people I've been spamming the Cookbook quest in Abyssea Misareaux for seals. I've gotten the base armor pieces from the Bastion (Legs) and Cruor (Head) NPC's and started the Magian trials on both. Feet base pieces come from Gold Pyxis in the zones where those jobs' other pieces are purchasable, meaning SAM feet base armor comes from Vunkerl. After about 12 or 13 repetitions of the cookbook quest (can be repeated in 3 to 4 minutes) I ended up with 3 WHM, 3 THF, 2 BST and 1 SAM. People seemed to think there are patterns, but I believe it's completely random. As a result I think SAM legs +1 are just a matter of time and perseverance. I have yet to see SAM head seals reported from a quest in Attohwa, but at least we know that's where they come form. Seals from quests are grouped into 4 jobs per quest regardless of zone and one equipment slot per Scars zone (Misa: legs, Atto: head, Vunk: feet). Killing NM's is going to give better results when doing it in a group. Jewels and such come from higher tier NM's in Scars areas, which means there's little chance of doing them solo.

Double Rainbow. What does it mean?
Well, the Unkai+1 (Samurai) pieces are good enough as is that they make new and more optimal setups possible. Both Store TP and Haste on head and leg armor pieces is something we clearly haven't had before and it makes for great TP pieces. Together with a 5 Store TP jump on the Magian weapon makes a 5-hit trivial to attain. So far I've been rocking Hachiman body and gloves, which is pretty good in low buff situations where accuracy isn't an issue.

At 464 delay you need a total of 64 Store TP to get 20 TP per hit. You can use this tool I made for other situations. Twenty is needed every swing to get 100 in 5 hits. You can check the math! Now any SAM that cares about strengthening their job will have or be working on maxing Store TP merits, which will put them at 35 Store TP from the Job Trait alone. Add a Rajas, which is almost standard issue nowadays and you're left with 24 Store TP to make up. Store TP Magian weapon now gives 10 Store TP and the head and legs Unkai+1 add up to another 10 and Brutal (which is both a good WS and TP piece) give 1 Store TP. This leaves 3 Store TP to complete the 5-hit. If you are aiming for the Unkai+2 (and don't care about the meantime) or have Usukane feet, you can skip the next part.

Completing the 5-hit
The easiest and simplest way to finish the 5-hit would be a Rose Strap, but there are other options like a Goading Belt, a pair of Ace's Sabatons, Askar Korazin, Hachiman Domaru, Hachiman Sun-ate. Some of these aren't very good at all, but I'm trying make a point. My point is that there are a lot of possible 5-hit builds depending on what's available to you. Also, don't forget to factor in how much Store TP you want to WS in. You can drop below 64 Store TP as long as you make up the difference in the 4 TP swings.

If NO gear is out of the question
When you're counting Store TP of your gear and are easily surpassing 64 because you are 1337 and have amazing gear, then I hate you! Your options to 5-hit are not limited to the Store TP Magian weapon. With Unkai+2 head/legs and Usukane body/feet you're looking at 28 Store TP from those slots alone. With another 11 you can even 5-hit a 450 delay weapon. However, if you're this decked out you would likely be interested in 5-hitting the "Occasionally Attacks Twice" Magian weapon, which is now so easy you don't even need Rajas.

How about a 437 GK? Masamune's low delay is amazing for it's melee DPS. It would be unfortunate to give up WS DPS by taking longer to build 100 TP. The same thing goes for the STR Great Katana. The crazy thing is that with the proper gear and possibly Carbonara the 81 Store TP needed to 5-hit it is actually already available. I'm going to refrain from doing the math on this extreme situation since it's likely to be refined in future updates and involves too many changes in stats which is likely to mess up the accuracy of my calculation.

Weapsonskill Damage Keitonotachi+1
A 5-hit is all well and good, but there ARE sacrifices in WS damage and accuracy and haste during the TP phase. Add to that the fact that the WS damage counterpart of Keitonotachi now has a generic boost of 10% to ALL weaponskills, then the 5-hit option for the average Samurai loses much of its shine. When fighting tough mobs the sacrifice of accuracy in giving up is costly. When fighting tough mobs while hasted by magic spells, sambas and Hasso the sacrifice of Haste on gear is even more costly. The Weaponskill damage +10% option is trivial to 6-hit, leaving room for as much accuracy and haste as you can find including the use of Pole Grip. When fighting tough mobs this weapon seems to take the lead.

Fire/STR Hyogugusarinotachi+1
It's considered the second best non-"relic" option after OAT Great Katana. I recently parsed it and it lost to a 5-hit Keito build in a capped acc low-buff situation. However, that comparisson doesn't really do it justice. A 6-hit build with this weapon (437 delay) should come at little sacrifice with the new Unkai+1 pieces.

That's all fine, but what should I pick?
Backed up with my trusty Damage Calculation spreadsheet and a hypothetical situation and guesstimations of pdif and gear stats I'd rank the easily obtainable weapons as follows:

6-hit STR Hyogugu+1 and 6-hit WS+10% Keito+1 are too close to call
Both allow for the same amount of Haste while the higher accuracy of the Keito build can let it outparse the STR weapon when you're uncapped. When your accuracy IS capped, the STR GK wins on paper.

The 5-hit Store TP Keito is the clear loser in any uncapped and buffed situation
Its sacrifice in Haste and accuracy are what do it in. Why then did my last entry said it won a parse. The short answer is: "I'm not entirely sure". The long answer is that the increased Meditate efficiency is very strong, but when plugging in my adjusted WS-Rates into my spreadsheet the difference is small but STILL in favor of the STR Great Katana. So, something in practice is different than my assumptions and estimations in the spreadsheet. I have yet to figure out exactly which factors are causing this relatively large discrepancy. I say "large" because even when messing around with my fSTR and pdif numbers the STR GK should never lose according to my own calculations. Perhaps the increased Meditate efficiency when I "discovered" it made me pay better attention to the timer as well as my gear swaps and general playstyle. This could have skewed it enough to make the GK win. I'm starting to see why people disapprove of parses as "evidence", but it's still worth confirming your hunches.

This is what the numbers vaguely look like. I'd like to add that I find any difference smaller than 2% to be too-close-to-call.

Sunday, September 5, 2010

Magian Katanas in practice

I believe it's been two weeks since I finished the Lv80 STR Great Katana. Since then I haven't had the chance to go out and compare it to the Store TP Keitonotach until last friday. My expectations going into this were that the STR GK would win by a small margin. The reason for this were the broad conclusions I made after calculating DoT in my blog post a while back. There the STR GK won the contest over a 5-hit Keito because of its higher melee DoT and ability to wear Dusk Gloves while maintining a 6-hit.

The conditions were as follows. I went out to Boyahda to help Kingcem with his trial. I started with the STR GK (Hyogugusarinotachi) wearing Hachiman Domaru and Haste in all the regular places (W.Turban, Swift, Byakko's, Dusk, Fuma). After about 30 kills (4 of which I removed from the parse below) I switched to Store TP Keito and swapped out Dusk Gloves for Hachiman Kote. Target mobs were Elder Goobbues. These were the results for the STR GK:

And this was for the sTP GK:

As you can tell, the Store TP Keito won the comparisson. There are some interfering factors like the Paralyze gaze attack slowing down Kingcem or me, but I believe it's overall a fair comparisson. So why did the sTP GK win? The first thing I noticed going from 6-hit to a 5-hit was how efficiently I could use Meditate. Since there is no room for lowered Store TP in this setup, I WS in the same (albeit it slightly higher) Store TP as I melee in. Each Meditate tic is 20 TP and each swing is 20 TP. When I'm 1 Tic or Swing away from 100 TP I spam my WS macro and WS at the first possible instant. "Wouldn't this mean you could get a slower swing if you do a melee swing while wearing your WS gear?". Good point! However, think about it this way. The extra 20 TP you have left over now that you didn't wait for a visual confirmation of 100% TP is worth one swing, which takes much longer than the bit of time you may or may not have sacrificed by swapping in your WS gear before reaching 100% TP. So, it's a net gain as far as I'm concerned. In addition to the increased Meditate efficiency, there is also the benefit of higher WS rate. Look at the Kparser performance tabs and you'll see the balance of Melee and WS DPS is shifted in favor of WS on the 5-hit build by "a lot". This is pretty obvious considering you need 1 less swing to WS.

This farming level of mobs is not the typical scenario where you'll find yourself worrying about DoT. We lacked any magic haste and I didnt' use Hasso at all since we were both tanking so much and had no healer. The sacrifice in Haste would weigh much heavier if we had other forms of Haste. This is another reason why the 5-hit build gets a boost. Since my average attack damage was around 111, one can not argue that my pdif was capped and that the STR GK's DoT suffered because of it.

The most important thing this experiment taught me was that I completely underestimated the effect of a 5-hit on WS rate. It's not just one less swing, it's also a considerable boost in Meditate efficiency. The most important question this experiment raises is: At what point is the 3% Haste sacrifice enough to make the STR GK take the lead.

An interesting thing I discovered was that Store TP affected my Double Attack procs on WS. Apparently this has been known for a long time, but I just never realised it. With 68 Store TP during WS I got a 22 TP return after a WS where Double Attack proc'd. With 68 Store TP my WS return is 20.4. Add to that an extra attack at 68 TP it's 1.6 added for a total of exactly 22.0 TP after WS.

Now if you were able to play perfectly, you could swap out gear pieces so that you are left with 60 Store TP total and still maintain a 5-hit, unfortunately it's not enough to simply remove Hachiman Kote, as you only have room to remove 4 sTP total, because

(100 - 22.0)/4 = 19.5

for which you still need 60 Store TP to achieve. Considering the relative infrequency of DA procs I'm not going to try and make such a situational gear-set. WS rate is THE determining factor for a SAM DoT calculation. Since Meditate influences this and its effect on WS rate is also dependant on what your WS rate was like BEFORE considering it, it's a difficult thing to account for. I'd like to repeat this comparisson in a situation with less down time and more buffs to see if the STR GK does indeed catch up and/or beat the 5-hit build.