Luck Of The Draw: A Corsair's Guide *NEW*

I wouldn't say I'm overconfident with spreadsheet numbers. Because if I do then I wouldn't go out and confirm the result again with parses.

There are also several occasions that I use builds different from highest spreadsheet numbers just because the situation favors something else.

For example, if build 1 has 3100 dps, 2.1 round/ws on spreadsheet, build 2 has 3000 dps, 1.9 round/ws on spreadsheet, then I'm going to use build 2 on fodder that dies in 2 ws because obviously build 2 will kill faster than 1 if both set kills it in 2 ws.

However I think spreadsheet as a tool is fast to get a base result then we can move forward with the number as a base. And certain data like ws avg and round/ws is convenient to find the most optimal builds.

And for the most part, the spreadsheet numbers aren't far off from real parses. I've never encounter a scenario that shows X build being 3000 dps ahead on spreadsheet and ended up 2000 dps behind in real parses, it doesn't happen.

In terms of arma v.s fomal, my opinion is that fomal generally wins because of SC and it doesn't need am3 management. However if everything lines up I have no doubt arma can win too. It's not the conclusion draw from spreadsheet I have, but everything.


If 1 person said X is better, then maybe it is confirmation bias.
If 2 person said X is better, then I would question it too.
If 3 person said X is better, I plug in the number to spreadsheet and sim, and got the result that X is better, then I would have solid evidence to believe it's true.

If one person pop out of no where claim Y is better after several sources favor X, at this point I would question other variables not considered rather than say every source that favored X are confirmation bias.

I would characterize you as overconfident in the spreadsheet results and in the amount validation that you claim you do. How often do you "go out and confirm the results with parses", and how controlled is it when you do? You literally just stated that their results should be used as a baseline, but we know that they aren't even close to accurate when it comes to RMEA comparisons (which people use them for regularly.)

I don't put a lot of faith into the conclusions you make. There have been several occasions in this topic where you have been presented with clear evidence that is contrary to your pre-existing beliefs and you simply recommit to whatever you said before regardless.

Which debate are you talking about? Most of the topics brought up here about dps difference does not have real "evidence" to prove otherwise either.

When we had arma v.s fomal debate who posted arma v.s fomal parses in controlled environment to support arma being ahead is wrong? Nobody.

When we had rostam b v.s a debate who posted parses to compare them? Nobody.

When we had Ra v.s melee leaden debate there was one parse, from someone without rostam B in melee set, got Ra being 1k ahead out of 470k dps, and not really considering dps lose from mode and weapon switch. I draw the conclusion completely based on that fact, pointed out that parse can't be a real evidence because above reasons and you called me biased.

If someone wants to post parses to prove spreadsheet wrong, go ahead. The problem is nobody do that then they proceed to call others biased if they use spreadsheet as a base. If evidence doesn't exist, what do you use as a base to determine what to use? We may as well play ffxi naked.

I feel this debate now turn into something like this:

A:Hey what's better?
B:X is better because spreadsheet say so.
C:Noooo spreadsheet is wrong, you are wrong.
B:proof?
C:You are so biased and overconfident with spreadsheet! You shouldn't use spreadsheet as a base! You are so wrong!
B: OK I'm wrong, evidence?
C:but you are just wrong. I don't believe you!
B: post your evidence then gdi.

If you want to prove spreadsheet wrong, post real evidence and people will change builds by themselves if the result is convincing enough. That's not the case for Ra v.s melee debate pages ago.

:]

Yup, took this opportunity to promote my sexual desire in ffxi :D

I'll try and make this as succinct as possible. You believe the spreadsheets are accurate enough for DPS comparisons although it's been demonstrated that in at least the case of Masa vs Doji, they are wildly inaccurate. These conclusions were drawn from parses on zerg fights and from prolonged parses on Neak. I don't have the parse data offhand although you can find the summaries if you search around (tbh I don't know which threads were posted anymore, but they are there.) Other players have drawn similar conclusions about the white damage ratio in the spreadsheets and have vocalized it on the forums. This is enough information to conclude that they are not valid to draw clear conclusions about DPS. You've ignored this and continue to believe they can still be used as a valid baseline. You demand that people need to give you evidence that the spreadsheets are wrong when that evidence has already been presented.

In regards to other scenarios where you've ignored evidence -

RA vs melee
Leaden vs savage

In both of these discussion you made assertions one being better than the other or 'not worth it' without concrete supporting arguments (as has been pointed out, spreadsheet result is not good enough) even when being presented with evidence to the contrary (some strong, some weak.)

I'd like to see methodology and data on this sam parse. Every sam I have seen has said masa > doji when ignoring skillchain dmg in zerg situations. It doesn't mean they are not wrong but that's a pretty big blind spot since any sam with masa should have aeonic at this point.

For the zerg fights, we prebuff with crooked XI rolls (Samurai Chaos Fighter's Miser's.) I don't remember all of the other buffs but we end up getting capped attack and accuracy. An addon is used to execute weaponskills the instant TP breaches 1000. We have not parsed and analyzed too many zerg fights, but in all of the Masamune parses we did the white/weapon skill ratio was never greater than 25%. The spreadsheet predicted a greater than 50% split.

For the Neak fights, we use BoG Frailty + entrust Fury for every fight and I believe either Samurai + Chaos was used, possibly Samurai + Fighter's. Addons are used to engage, to weapon skill and to apply debuffs. The only manual input is buffs between fights. It's very mechanical. The white damage split results were the same and Masa never outparsed Doji.

what gear were they wearing? I just find it extremely hard to believe every sam with both hasnt noticed doji > masa. I never bothered finishing masa on mine with the gil pit that is rostams

In terms of leaden v.s savage in wave 3 on fodder, I have multiple parse result for both ws avg to draw conclusion from. I have done them with both, in pt with malaise and without. It was not based on spreadsheet at all.

And I already made the conclusion that without surplus of mage buffs, it's not worth it. If you do, then it outperforms savage slightly. How is this conclusion biased? This conclusion is agreed by Comeatmebro and Chiaia in the discussion too.

On the other hand, you favor leaden from the get go because you used ranged strat to begin with, made a claim that leaden hit 60k in your setup, then back track your argument that 60k is only doable with gambit Rayke both up. To Me that's biased to completely ignore the dps with normal buffs and only use the number with max buffs to begin with.

And yet you are here saying I'm biased but not you, ok.

In terms of ranged v.s melee, I didn't use spreadsheet to compare them. Spreadsheet doesn't factor ws delay for ranged, so ranged dps often come out much lower on parse. I simply question ranged being better way to spam leaden in wave 3 in a melee pt.

Note that there's huge lag with ranged attack in dyna, but melee not so much. So I'm not convince ranged being by all end all best dps style in dyna if not for lags.

Pardon my intrusion but... Isn't a spreadsheet or a simulator a measure of an absolutely perfect situation inside a vacuum?

It's cool and all to have calculated numbers on what's the penultimate gear/ability/weaponskill combinations, but isn't it theoretically impossible to reach said numbers in-game due to limitations like human reaction time, human error or circumstances out of your control like Lag?

Sorry, carry on.

This wasn't even spreadsheet argument at this point. I made a point based on real parses info, then snaps pop on the forum claiming my points are wrong because it's different from his parse result.

So yeah, only snaps parse results are valid, everyone else's are biased and wrong if they are different.

Same. Masa being ahead isn't just spreadsheet result, but also Austars sim which doesn't favor white dmg as much.

In fact masa is even further ahead than doji on SIM because masa favors zanhasso build(with tp overflow) if I remember correctly.

So if someone noticed doji being ahead in controlled environment, I would question gear difference. What feet are you using in masa tp set?

Of course but that's a silly argument. The argument isn't spreadsheet number are not what real player would do. The spreadsheet doesn't model a ton of the game and FFXI mechanics down to the minutiae get fairly complex on ***like post ws/ja delay on attack queue.

The argument is if the spreadsheet is accurate enough to produce what is best gear to wear for various situations. FFXI lacking in a training dummy makes any kind of validation difficult.

Some kind of robust sim would be a step up in transparency and precision, but also built on a lot of assumptions and testing because ffxi is convoluted as ***.

We used whatever was in the guide at the time. I don't think people 'notice' because humans are not machines that rigorously test and compare everything they do. Unless there is a superior choice and it's clearly evident (Weber's law), they either parrot what they hear others say or they lean toward's their personal preference. Masamune (and Samurai in general) got put on a pedestal due to some popular posters with no supporting evidence that it was actually better than the alternatives.


There was no backtracking. I provided evidence of the numbers and stated the buffs that were used. I don't think you need to overload your party with mage buffs to tilt the balance toward's Leaden either. For Volte Leaders, I think the only things you need are Rayke, some source of magic evasion reduction (Dark Threnody II is what we use) and gear with reasonable amounts of magic accuracy on it. For volte trash mobs, you don't need as much magic evasion although having Rayke on those is not a realistic expectation. That said you can still get some good numbers.


This was from our last run. We did have bolster malaise up for some of those mobs although you can get similarly good results just using BoG + EA.

I never stated that ranged was better than melee either, or 'by all end all dps style' as you put it. I stated that DPS while under the affect of Triple Shot was likely higher than melee. Someone parsed it and got this outcome (not a very controlled test), and I did some very controlled experiments on brittle rocks to further analyze it.

In regards to ranged exclusive in dynamis, I feel it's practical because you don't need any additional defensive support or defensive buffs. It's easy to semi automate and the damage is good enough to clear the content with plenty of time remaining. It's easy and reliable. Our group uses it for these reasons, not because we think it's the best DPS option.

I'm not actually making this argument. I think the spreadsheets are good enough to select weaponskill and to some degree melee/ranged gear but otherwise aren't useful or valid for comparing different weapons and playstyles.

The reason why I said I'm not biased toward leaden on wave 3 because last run I did with someone using bis leaden set + malaise, the ws avg from entire run was less than 35k, which is barely better than savage. Some of the ws was as low as 2k on leader.

So Yes, I have number to back up my claim, I didn't make assumptions out of nowhere like you said.

Yes, you can post a screen shot with more buffs, with perfect setup build FOR leaden, with rayke. That doesn't mean my claim about leadens performance in melee pt is wrong, nor biased. It's simply different result under different conditions. And what matters is avg ws in entire run, not spike when Rayke is up.

Afa: hey guys, what's the best way to DD in dyna?
Snaps: leaden because we can do 60k with it.
Afa: well I'm not getting this kind of number, we do wave 3 with melee setup.
Snaps: you are not trying hard enough! You are biased!

It's emotional argument at this point.

People are terrible at estimations, which is obvious. Its why I am asking for evidence.
From what you claimed the difference between masa and doji was large. I actually find this way more interesting than anything about cor. Cor has a huge margin of player skill/williness to bot on the effectiveness of job. I think a dipping bird can play sam to 95% ability.


Well, I wasn't quoting you so not sure why you think that I was saying. After messing with the current spreadsheet I'd say that's a somewhat fair assessment, but also consider that there isn't much of a better method. FFXI has no training dummy.

YOU'RE TEARING THE FAMILY APART MOM AND DAD

At least these argument are somewhat substantive compared to
"OmG tElL mE aGi AnD mAb 2 oNe ShOt StAtUe"

My opinion on spreadsheet is that people uses "spreadsheet is wrong" as an excuse way too often in gear discussion, while none provides better answer to solve the problem.

There are so many gears in ffxi, and there are countless combinations when it comes to what gears to use. If I'm going to find the best set in X condition, the fastest way is to use a spreadsheet.

If you don't use it, then the alternative way would be:
1) parse in a controlled environment.
2) do math manually
3) guesstimate.
4) don't care, pt naked.

We can all agree that 3) is the worst. So what's left is 1) and 2).


1) is not entirely reasonable if my goal is just to find answers for minor things like epona v.s chirich +1. I'm not going to go out, parse for 3 hours with epona and get my dps in the long run, swap to chirich+1 then parse for 3 hr again just because I can't decide what ring to use in a 15 yr old mmo. That's the worst use of time ever.

2) is another way to get the answer. It's faster than 1) but slower than spreadsheet. You can get numbers like ws avg, round/ws, avg hit etc and figure out possible dps using known formula.

That being said, spreadsheet are already based on these known formula. Unless there's glitch somewhere, the avg hit, ws avg data you can get by calculating manually will be mostly identical to spreadsheet results unless the sheet that you use has glitch unfixed somewhere.

Which CAN happen, but I don't think the difference between spreadsheet result and real parse is caused by spreadsheet glitch. I think the difference between theory craft dps and real dps is caused by unknown variables in real ffxi. And the only way to get rid of variables is to do parse in a controlled environment. Do math manually and slowly will not solve this issue.

Then it goes back to the downside of 1). With so many gears nobody has that kind of time to figure out best set for all situations. We will have to settle for theorycraft dps 90% of time for majority of gears.

If someone wants to post parse in controlled environment to prove spreadsheet or any theorycraft dps being wrong, it is a good thing and it should be encouraged. In fact, I believe I was the only person who posted parses in arma v.s fomal debate, lol.

But if someone just want to pop on the forum judge people's opinion with "but spreadsheet is wrong so you are wrong" or "you are biased" then it's nothing but emotional argument. It doesn't solve problems.

So yeah, I use spreadsheet as a base, not because I think it's by all end all the best and only way to find out the best gears, but because every other way are even less realistic and we have to settle on something somewhere so we can play the game.

If you don't like spreadsheet, feel free to parse for every set in a controlled environment and waste time, or do math manually and get similar result as spreadsheet, or guesstimate, or don't wear gears. Whatever. It's your $13 not mine, so I don't care what people wear in game.


If I remember correctly, the conclusion of masa being better came after Austars sim which shows masa pulled ahead with zanhasso build, which caused quite a bit of tp flow and made tp bonus from doji less relevant. But I can remember wrong, it's been a while.

So if that doji v.s masa parse was done before the conclusion was made and they were not using the right feet, the parse result may favor doji.

That being said, I don't think that would be the reason why spreadsheet is inaccurate. The ws avg and avg hit per round from spreadsheet is still based on math and formula. If there is a gap, the variable probably came from somewhere else.

I wouldn't draw the conclusion of all answer you get from spreadsheet is incorrect just because doji out parse masa with certain gears in snaps pt. That's a pretty big jump.

Agreed. I tend to do (1) and (2) myself. But then again, I'm only concerned with average per hit damage on melee attacks and average Weaponskill damage. The rest is minutia that falls within a 2~3% difference usually dependent on circumstances I mentioned previously.

Is there a around about figure for what each % of WSD is worth in MAB for a piece of gear?

The value is pretty fluid based on current values of gear and buffs. using the calculator is best bet:
https://www.ffxiah.com/forum/topic/52018/luck-of-the-draw-a-corsairs-guide-new/28#3387371
For an extremely rough estimate 3 to 4 mab for 1 wsdmg?

Have no rostam, but I'm surprised you all go/want to go with augment path A & B for rostam without considering the path C.

Isn't the phantom roll effect +8 (so +3 over the bartaria ring / +1 over the regal necklace) more beneficial to you and your group than few extra WS dmg?

Maybe indeed if you get the necklace, the +1 effect isn't so gorgeous ? But for someone with just the ring,the +3 effect should have some significant appeal, no?
Just wondering how to *** the benefit of the phantom roll effect +

https://www.bg-wiki.com/bg/Category:Phantom_Roll has list of values for +1 roll. For most common rolls, it adds 3.12% attack to chaos and 4 stp to sam roll. If you don't have regal neck then path C is massive upgrade. If you do its less of a big deal but it does have the draw of easy 100% crooked up time.

There are 2 reasons its not super hyped:
1. Requires to swap out mainhand to roll which is annoying.
2. It doesnt add near as much personal dmg as A or B.

If you only play cor in group content and especially lolsmn nonsense then C is prob best path for you. If you solo/lowman on cor a lot its a pretty meh aug.

Personally, I went Path C first. To me, COR is *always* support first, DD second. The +1 phantom roll over regal is minimal, but useful. But most important to me is the +60 second roll duration. It means it's easily possible to achieve 10 minute rolls with comfortable group 2 merits. And with that, it means you can achieve perfect crooked cards rotation.
Next up for me is probably the Store TP path. It's not BIS for melee (follow up attack is), but it's also not far off. And it works for both ranged TP and melee TP.

I’m curious if compensator now becomes a necessary phantom roll item since you might start swapping rostams for its benefits anyway

CC rotation was already doable with compensator. Able to change the merit for snake eye 5/5 is the only benefit extra duration has, IMO.

As far as 3% attack 4stp goes......I'm often capped attack in ambuscade so it's really 4stp boost lol.

It's a matter of preference, but I wouldn't let go of my snake eye 5/5, even for that.