Salvage - Bhaflau Remnants - thread2

[divisortheory]

Their drop rate seems quite high. Everyone I've known who has seen more than 5 or so pops of anything always has something to show for it.

We're 6/~25 in the zone including 4/9 on the bees (in ~17 Bee Rampart attempts, we've seen bees only twice, a 4pop and a 5pop) which isn't a large sample to be sure. Pretty much everyone is reporting similar drops though, which seems to be significantly different from the ~10% in Arrapago. Overall though the L35s per run (as opposed to per NM) have been about the same for us in the 2 zones, simply because of the number of Rampart runs we've done without seeing an NM at all...

(oh, and I don't get the Dynamis analogy... what there doesn't seem to have a "normal" distribution of pops or drops?)

You've never been to those dynamis runs where you get like 18 AFs and 4 100s in the same run? There have been times in Dynamis where I've seen 1 or more AFs in the treasure pool almost the entire run. The drop rate just doesn't seem constant across runs, similar to how the pop rate of Rampart HQs doesn't seem constant across runs in Bhaflau.

[Ambir]

Their drop rate seems quite high. Everyone I've known who has seen more than 5 or so pops of anything always has something to show for it.

We're 6/~25 in the zone including 4/9 on the bees (in ~17 Bee Rampart attempts, we've seen bees only twice, a 4pop and a 5pop) which isn't a large sample to be sure. Pretty much everyone is reporting similar drops though, which seems to be significantly different from the ~10% in Arrapago. Overall though the L35s per run (as opposed to per NM) have been about the same for us in the 2 zones, simply because of the number of Rampart runs we've done without seeing an NM at all...

(oh, and I don't get the Dynamis analogy... what there doesn't seem to have a "normal" distribution of pops or drops?)

You've never been to those dynamis runs where you get like 18 AFs and 4 100s in the same run? There have been times in Dynamis where I've seen 1 or more AFs in the treasure pool almost the entire run. The drop rate just doesn't seem constant across runs, similar to how the pop rate of Rampart HQs doesn't seem constant across runs in Bhaflau.

Well yeah, but that's not random, that's still a normal distribution. For a statistical distribution the most common result is the Mode value. Values close to the mode have a higher probablility than those further away, and the further from the mode value a result is, the lower the probablility of it occurring is. You get a nice bell shaped probability curve.

For Dynamis the 2 shells I'm in (long established shells) are pretty consistent in what they kill each run. Both average around 10 AF 2 in starters and ~4AF2 in Xarca. We can get as high as 20 or as low as 3 in starters, but it's uncommon. And the further away from 10 the number, the less likely the result seems to occur Likewise in Xarca we can get 0 and have had as high as 10, but 3-5 is pretty normal. Our results (and both shells have ben predominantly Xarca for 18 months now) are basically a normal distribution with an average and some higher/lower.

For Bhaflau Ramparts, there are a lot of times (over 50% ) where we see no NMs. If the chances of an NM popping were equal at all times were equal, then getting a single NM only should be the next most probable result, and it isn't. Nearly everytime someone gets one NM, they get 3-6 NMs and 1NM only seems really rare. That says to me that this isn't a normal distribution, and that sometime ramparts the chances of an NM popping are very low (quite possibly 0) and that some ramparts the chances are significantly higher, probably around 3-4% per pop (judging from the reports of multiple pops here).

The real question is why are all ramparts not equal, and is there anything we can do the influence or predict which ramparts will pop NMs?

[Kaylia]

I'm with Ambir on this one. I know bad seeding can give stupid result, but I don't think it's case here (or anywhere in the game). I believe the rampart have an "on" and "off" mode. When nothing pop, it's off and nothing will ever pop. When it's "on", there could be a 5% chance to get NM, which would explain the 2-9 NM pop people have seen. Something like this would explain both extreme.

Is there a controllable condition to make the rampart "on"...no idea. Could be related to the method of pulling it? Zhayolm rampart have weird reaction to magic after all. Pull with element of the day? Maybe there is a single one flagged every run and other 3 are off? Or maybe it's just a 2nd random % that determine if they are on or not.

[divisortheory]

Let's assume a fixed probability of N, and that the rampart spawns 75 mobs (is this right?)

After 75 mobs, we have the following probabilities

N = 10%

P(0) = 0.037%
P(2) = 1.27%
P(5) = 10.8%
P(7) = 15.45%
P(10) = 8.8%
P(12) = 3.4%

N = 5%
P(0) = 2.13%
P(2) = 16.4%
P(5) = 14.9%
P(7) = 4.7%
P(10) = .29%
P(12) = .03%

Since we know from experience that P(0) is more like 25%, we can extrapolate and determine that if the pop's are generated according to a normal (i.e. perfectly random) distribution, then N would have to be something like 2%. But that makes everything else ridiculously out of whack as well. For example, if N=2%, P(2) = 25% and P(12) = .0000029966813%.

Conclusion: It isn't a normal distribution and there is another factor involved.

I don't see any reason why the idea that day/moon/voodoo affects the pop rate is any more compelling than saying that it's seeded when you enter. Both situations could easily lead to the observed results. Seed it with 1% sometimes, and 10% other times and you end up exactly with the observed results. And for me anyway, Occam's Razor always decides the winner between 2 baseless theories that both fit observed results.

[Francisco II]

We've only attempted the Skirmish Pephredo rampart once, and in that one attempt the Rampart popped three bees. The third and final bee dropped Hikazu Sune-ate.

I'm pretty sure that it was Earthsday, around 15:00-19:00 maybe... RDM aggroed the Rampart with Blind.

[Ambir]

Only that very little if anything in the game is truly random. Drops almost always follow a binomial distribution. Here, two effects appear to be involved, a rampart with a binomial distribution based around a %pop chance of maybe 3-5% and a rampart that pops with nearly a 0% chance of popping an NM (even if there was a 0.5% chance of popping a NM each pop, there would still be around a 30% chance of getting a single NM from a run and noone is seeing that).

If Occums razor explained all the drop/pop rates, then the frogs in Zhayolm would have just been assumed random pops before their pop conditions were known. They clearly aren't and in a lot of cases we can now control their pops. There's a good chance that there is a condition controlling whether a rampart pops or not. Whether it's knowable or not is unclear, but very little in Salvage is completely random, and if it can be predicted, the efficiency in this zone would increase by a huge amount.

[divisortheory]

Here, two effects appear to be involved, a rampart with a binomial distribution based around a %pop chance of maybe 3-5%

I just proved that it wasn't based on a normal distribution >.> I know BD != ND, but given a large enough sample size, which obviously we are assuming since we're discussing the true drop rates, they are almost equivalent.

even if there was a 0.5% chance of popping a NM each pop, there would still be around a 30% chance of getting a single NM from a run and noone is seeing that.

Actually there would be about a 3.6% chance of getting exactly 1 pop, and about a 3.7% chance of getting at least 1 pop.

[Ambir]

I'd been assuming 100 pops per rampart, but my figures give (given each pop was 0.5% chance of popping an NM)

Probability 0 pops:
= .995^100
= 60.6%

Probability exactly 1 pop:
= 100!/(1! x 99!) x 0.005^1 x 0.995^99
= 30.4%

which would mean that if all the ramparts that were popping 0 NMs really have a low probability of popping a NM (say 0.5% each time) then we should be seeing a lot of ramparts popping 1 only NM.

Our shell has had 15 runs alone on the second floor rampart pop 0 NMs, 2 pop 4 or 5 NMs and nothing in between. Tha fact that we are seeing almost no ramparts pop 1NM indicates to me that a most of the ramparts that are popping 0NMs aren't actually capable of popping a NM, and the it isn't just bad luck pop after pop.

(It's been awhile since I did binomial theory, but I think my calculation is correct...)

[divisortheory]

Ahh, yea I was assuming 75 pops (because it definitely pops enemies much slower than 1 mob / % of HP) for one thing, and also I added an extra 0 accidentally and did .05% instead of .5%.

So with 75 NMs, P(0)=68.7%, P(1)=25.9%, and P(>=1)=31.3%. Nonetheless, if we extrapolate this out we get:

P(0)=68.7%
P(1)=25.9%
P(2)=4.8%

So P(>=3) = .6%? I think not.

Regardless of how many people have seen it pop 1 NM vs. 2 or more (I think I've seen 1 Gate Widow, but who knows it's been a while since we did that one), it's pointless really talking about it because arguments can easily be made to say "I think drop rate is based off a roll when you enter the zone", or "I think it's either on or off", or any number of other theories.

Here's a new theory. The probability that the next mob will be an HQ spawn depends on whether or not the previous mob was an HQ spawn. It doesn't even matter if it's true, because we will never get any closer to figuring out how to get it to pop more, which is the only important thing, because that's how it was designed. Even if we figure out the exact formula for determining the pop rate, it will have been designed in such a way that factors which we cannot influence are involved in the calculation. Like crafting, for example. Everyone always talks about how crafting is very streakish, just like how you get 3 HQs in a row, or NM drops like how KB drops 3 D. Rings in the same month, just like how Fafnir drops 5 Ridills in 2 weeks. That shit has been in the game for 5 years and nobody (who isn't full of shit) has figured out some voodoo secret to getting the expected results, why must people keep clinging onto the idea that this will be any different?

Keep fighting the rampart, eventually shit will pop. If it was possible to get it to pop more consistently, the zone would be broken because people would be walking out of every single Bhaflau run with 3 or more Lv. 35s. Low pop rate, high drop rate (Bhaflau) vs. High pop rate, low drop rate (Silver Sea). Sounds pretty balanced IMO.

[Ambir]

Regardless of how many people have seen it pop 1 NM vs. 2 or more (I think I've seen 1 Gate Widow, but who knows it's been a while since we did that one), it's pointless really talking about it because arguments can easily be made to say "I think drop rate is based off a roll when you enter the zone", or "I think it's either on or off", or any number of other theories.

Here's a new theory. The probability that the next mob will be an HQ spawn depends on whether or not the previous mob was an HQ spawn. It doesn't even matter if it's true, because we will never get any closer to figuring out how to get it to pop more, which is the only important thing, because that's how it was designed. Even if we figure out the exact formula for determining the pop rate, it will have been designed in such a way that factors which we cannot influence are involved in the calculation. Like crafting, for example. Everyone always talks about how crafting is very streakish, just like how you get 3 HQs in a row, or NM drops like how KB drops 3 D. Rings in the same month, just like how Fafnir drops 5 Ridills in 2 weeks. That shit has been in the game for 5 years and nobody (who isn't full of shit) has figured out some voodoo secret to getting the expected results, why must people keep clinging onto the idea that this will be any different?

Keep fighting the rampart, eventually shit will pop. If it was possible to get it to pop more consistently, the zone would be broken because people would be walking out of every single Bhaflau run with 3 or more Lv. 35s. Low pop rate, high drop rate (Bhaflau) vs. High pop rate, low drop rate (Silver Sea). Sounds pretty balanced IMO.

And people said keep running through Bhaflau and the frogs will eventually pop, they're random (turns out lots are time based).
People said the Bhaflau 4th floor N rampart that warps you to the next floor is random (turns out to be related to the day you enter the Salvage).

If there is a state of ramparts that cannot spawn an NM, then that information can be used to increase your drop chances during a run, even if we can't predict the cause. If the chances of a NM popping are 4% when the rampart is active (as opposed to when it is a 0% chance of NMs popping) there's a nearly 90% chance that if no NMs pop in the first 50 pops, then no NMs will pop from the rampart at all. If true, it means that you can cut your losses on an inactive rampart and try another.

The cause may well be random. If it isn't though, someone will stumble on the trigger eventually, and the more people are looking, the faster that will be. And after the number of runs through Bhaflau I've lead with no NMs, let alone no drops, I think it's definitely worth keeping an eye out.

I wouldn't complain if a zone was broken in my favour for a change.

[aurik]

And yet the chance of getting a NM could also depend on how many mobs are out at a given time, or based on some randomly chosen seed, or a million other things. There's a huge chance that we'll never figure out how to influence the number of NM popped per rampart, and an even larger chance that if we do stumble on some method we'll never be able to prove it to a level of significance where I'd be confident in it.

[Francisco II]

Run A:

Went in with six, it was mostly an experimental pickup group...

Cleared most of the first floor, west path in usual fashion, Mad Bomber was up, we attempted to kill it and failed.... Apparently some people's idea of "going all out on damage" is popping a Reraise Hairpin with the mob at 50%. Not one, but TWO people did this (a MNK and a WAR).

On the second floor, I made it clear that we did not want to use any magic damage against the flans. I specifically mentioned Formless Strikes. We began killing the Flans counterclockwise. On the third Flan, the MNK exclaimed, "Wow, Formless Strikes makes me do like 10x damage!" Ugh... I finally got everyone to disengage with the mob at 4%... our RDM slept it, they all logged out... I kited it until I lost hate. Fortunately, it regened and we were able to get it to dawn spikes again.

We popped the rampart, and the rampart spawned three Skirmish Pephredo... one of which dropped Hikazu Sune-ate. Noteable behavior: RDM aggroed the mob with Blind. I was somewhat frequently using Spinning Attack, if we had several mobs with relatively low HP.

Run B:

Cleared the first floor, west side, as stated above.

Mad Bomber wasn't up on our side.

Killed the Flans clockwise this time, no magic incidents.

The Rampart did not spawn any NMs.

[Sechs]

I have many issues I would like to compare with the opinions/suggestions of players with way more experience than me on this fight. For people who do not know me, I usually go to Salvage on MNK. This first post is going to be pretty long because I have a lot questions, and I have problems in trying to be brief, also... I am a total noob. Please try to bear with me.


1) Why MNK/NIN?
I've always seen MNK/NIN being used to dual-tank LBC. Why? I Can see it over a PLD/NIN (who would give more TP, do less damage, and could have problems to keep hate over another DDtanker), but why over a NIN/WAR for example? While MNK/NIN can only use WS as a "tool" to exchange hate between them when needed (no Provoke), the hate exchange thing could get a bit too tricky with 1 MNK/NIN and 1 NIN/WAR, over the situation where you have 2 MNK/NIN. But what about 2 NIN/WAR? Wouldn't 2 NIN/WAR tank better than 2 MNK/NIN?

2) Which food?
It's been my first try with this group, all has been very messy and we barely brought him down to 60%, with many errors from our side. I didn't notice an insane eva from his side, but it's true I haven't paid enough attention maybe (no, we didn't have a THF with Faint). Which food do you suggest? I was thinking about Carbonara. Otherwise Hedgehog Pie could be a nice choice (slightly boosts your Acc too). Suggestions?

3) Which gear?
Should we go all DD? All eva? Something in between? I went with my DD gear which you can see right below, only difference is I used Osode in place of Togi (I had VIT, STR and CHR unlocked, all equip slots except ranged). Destroyers, Walahra Turban, Kiriin's Osode, Byakko's Haidate, Bandomusha Kote, Dune Boots, Jaeger Ring, Rajas Ring, Ethereal Earring, Brutal Earring, Black Belt, Forager's Mantle, Faith Torque. Or should I go eva? Other gear pieces I have are:
Arakan Samue, Evasion Torque, Fuma sune-ate, Melee Gloves, Optical Hat o.O Well I think this last one would be a bad choice? I guess I could buy a pair of Hades's Sainti too, might help with Subtle Blow, Evasion and enmity, since I have Enmity-4 Merits.
What's the setup you use? Which one you suggest me to use?

4) Which cells?
We did some mistake with cells, next time we will act differently. What's your suggestions on Stats cells? Should we bother about them for MNK/NINs? Or isn't it worth the trouble and it's better to concentrate on equip cells?

5) MNK/NIN general strategy?
Basically, how does the fight "flows"? We cannot exchange hate easily, so I guess we need 2 MNK/NINs with similar gear/enmity, let them save TP to use when someone need the other to get hate, so he can easily recast shadows. Is this how things go? If everyone just uses TP as soon as they reach 100, how are we controlling hate swap the way we want? And what about a third MNK/NIN or a NIN/WAR staying far from LBC, using Chi Blast or Provoke/Enfeebles and being always ready to engage LBC in case one MNK/NIN dies?
In theory this seems a good failsafe in case one MNK/NIN dies. In practics I fear that the third member suited for this role would have serious problems to get hate once one MNK/NIN dies, and in the end the other MNK/NIN would end up tanking and dieing. Suggestions?
I like this strategy, but basically it seems to me that if one MNK/NIN dies... it's all lost.

6) General Suggestions
Anything you would like to share with me?
Gear? Macros? Useful jobs to support MNK/NIN? (we had BRD, but without instruments lol, no THF for Feint but we can get a pair, and absolutely no DRG). What about the Homing Missile gimping strategy? That way Bind would become a minor issue compared to before, and WHMs could concentrate more on healing and less on Eraseing.

Once again sorry for this long post, thanks for your precious time you wasted in reading this sh*t.

[Wizerd]

5) MNK/NIN general strategy?
Basically, how does the fight "flows"? We cannot exchange hate easily, so I guess we need 2 MNK/NINs with similar gear/enmity, let them save TP to use when someone need the other to get hate, so he can easily recast shadows. Is this how things go? If everyone just uses TP as soon as they reach 100, how are we controlling hate swap the way we want? And what about a third MNK/NIN or a NIN/WAR staying far from LBC, using Chi Blast or Provoke/Enfeebles and being always ready to engage LBC in case one MNK/NIN dies?
In theory this seems a good failsafe in case one MNK/NIN dies. In practics I fear that the third member suited for this role would have serious problems to get hate once one MNK/NIN dies, and in the end the other MNK/NIN would end up tanking and dieing. Suggestions?
I like this strategy, but basically it seems to me that if one MNK/NIN dies... it's all lost.

I don't feel like answering all your questions, as they are answered elsewhere, but I will respond to the bold above:

If your Monks have imbalanced hate, that's fine. Chances are that will be the case with any group, anyway. Hate will swap easily once the MNK who has hate starts getting hit, and that's his chance to put shadows back up if it's not possible when getting swung at. Your Monks WILL get hit, no doubt about that.

[TB]

1) Why MNK/NIN?

Yes, 2 NIN would tank better, but 2 MNK do more damage during the previous floor and on boss as well. You also don't need 2 MNK. I've been solo tanking most of our LBC due to us usually having only 1 WHM and 1 RDM.

2) Which food?

I use a Hedgehog Pie for a complete Salvage run. On Boss I get March, Minuet x2 and Madrigal and usually don't miss much.

3) Which gear?

I'm using almost completely my meripo gear with minor changes.

Destroyers (Hades Sainti IF a HM landed)
Olibanum Sachet (Tiphia for Asuran)
Turban,Peacock Charm,Ethereal Earring, Brutal Earring
Osode, Melee Gloves, Jeager Ring, Rajas Ring
Cerberus Mantle, Black Belt, Haidate, Fumas

This will max your Utsusemi recast with only 1 March and Haste and also add a good amount of Subtle Blow.

4) Which cells?

I lot all cells but CHR. INT and MND after RDM and BLM of course.

5) MNK/NIN general strategy?

Be prepared to cast Utsusemi alot. A completely enfeebled LBC can usually be tanked with ichi only, so you can save Ni for TP moves that strip shadows. With Slow2 and Elegy you can almost cast a complete ichi in between attacks. Save your Body Boost and other attack/accuracy up drinks for the last 25%.

[Sechs]

Guess I will then buy a pair of hades (with enmity-4 from merits, all those stats will help me) and equip Bibiki Shell, Osode, Haidate, Melee Gloves (+subtleblow, no -eva over Bandomusha), Faith Torque, Fuma Sune-ate, Forager, Walahra. 25% gear haste, def and stats from osode, no -eva, some +agi, fair amount of +att and +acc too.

TB a few more quick questions. Was your group using the Homing Missile nerf strategy? (killing 10 gears on 4th floor) How were you handling the tanking? On one side, being sure to have one mage Erasing Bind after Inertia Stream etc?
How did you guys do on Stats cells? Because on our next run we wanted to farm just a few stats monsters, and concentrate on Equip gear.
This is our tentative plan:
1st fl) West path, Kill bifron, trolls, Wamoura, then the 5 wamouracampa in last room, close to the teleporter, then up
2nd fl) kill 4 flans, then east path, use socket if present, then south, kill a few scorps/trolls for stats (we killed the whole room last time, it seems they only drop a certain a mount of stats cells tho, not all), then sneak/invi to next room and kill all for gear slots.
3rd fl) Kill some flans if someone still need cells then up
4th fl) Kill all 10 gears (we just fleeed to 5th floor last time)
We ended up with a bit more than 25mins at LBC, I think 10 mins is enough, so we can farm some more Wamouracampa on 1st floor and kill all gears in 4th, 15 mins for this should be enough.

All of this because we arrived to the conclusion that StatsCells aren't that relevant, except maybe some INT for blms or CHR for BRD. I would love some VIT, STR, DEX, AGI but I can live without, Gear is more relevant. At least... that's the conclusion me and my group arrived to. Maybe we're wrong? (we had 13 people btw)


Thanks for your sharing of knowledge

[Almalexia]

I would love some VIT, STR, DEX, AGI but I can live without

Goodness, no. Get yourself a pannus. -30STR makes an apparent difference in damage and, by extension, battle time on LBC.

[alphacat]

What would an ideal low man group look like for the boss? From what I hear any melee over 2 mnk is unwelcomed. So I'm guessing setup of mnkx2, rdm, whm, brd, blm is the minimum to do this boss with?

[Sechs]

I would love some VIT, STR, DEX, AGI but I can live without

Goodness, no. Get yourself a pannus. -30STR makes an apparent difference in damage and, by extension, battle time on LBC.

I had a Pannus and a VIT one last time. But the thing is... it's so random that you can't choose which cells to get. And if we need to kill all 10 Archaic Gears on floor 4, we need to make a choice... we can't get everything of everything. We either leave some slot in gear, or in stats... we can't get everything for every single thing for 13 members, and then have the time to kill 10 gears and at least 10 mins for LBC.
At least that's how things are for my group, for now. Maybe we'll get even faster with next runs.

@Alphacat
Yes... the more melee = the more TP, we all know the rest.
I think maybe a SAM could just meditate and do some random WS now and then, staying far for the rest of the fight, some DRG/SAM could do more or less the same and help with Angon.
THF I dunno... Faint would be useful, but for the rest I have no idea.
On one hand you would want only 2 MNK/NINs and a shitload of BRD/RDM/WHMs. On the other hand you know that with such a setup it's gonna be a pain to get all those praecipitatio cells without a socket, and it's gonna be hard to get up the floors when you start, with so few melees.
At least, that's how I see things, maybe I'm missing something.

[Vistani]

I would love some VIT, STR, DEX, AGI but I can live without

Goodness, no. Get yourself a pannus. -30STR makes an apparent difference in damage and, by extension, battle time on LBC.

I had a Pannus and a VIT one last time. But the thing is... it's so random that you can't choose which cells to get. And if we need to kill all 10 Archaic Gears on floor 4, we need to make a choice... we can't get everything of everything. We either leave some slot in gear, or in stats... we can't get everything for every single thing for 13 members, and then have the time to kill 10 gears and at least 10 mins for LBC.
At least that's how things are for my group, for now. Maybe we'll get even faster with next runs.

@Alphacat
Yes... the more melee = the more TP, we all know the rest.
I think maybe a SAM could just meditate and do some random WS now and then, staying far for the rest of the fight, some DRG/SAM could do more or less the same and help with Angon.
THF I dunno... Faint would be useful, but for the rest I have no idea.
On one hand you would want only 2 MNK/NINs and a shitload of BRD/RDM/WHMs. On the other hand you know that with such a setup it's gonna be a pain to get all those praecipitatio cells without a socket, and it's gonna be hard to get up the floors when you start, with so few melees.
At least, that's how I see things, maybe I'm missing something.

I don't think your thieves passing out will help anybody.