Anyone tried i90 ACN?

[T0rin]

I'm looking forward to taking teh time to write up a reply later this evening. I simply assume loldrg is a troll, but I'll prove him wrong regardless. Same bad arguments we've heard 20 times over.

[LateRegi]

I'm looking forward to taking teh time to write up a reply later this evening. I simply assume loldrg is a troll, but I'll prove him wrong regardless. Same bad arguments we've heard 20 times over.

So far his arguments have been more compelling than yours, and he clearly has a firm grasp of what he's arguing. Labeling him a troll is the most severe cop out possible, but you're likely not capable of producing a better argument than he, so I guess your "troll" response was expected.

[loldrg]

Same bad arguments we've heard 20 times over.

Oh the internet. Where we are expected to respect random person X's data like it's just been published in Nature and is about to win the Nobel Prize in Chemistry without actually being able to question the results on the basis of the experiment or demand verification/replication of the results by a 3rd party. And if someone questions the self-proclaimed expert without their own equally spurious data they are automatically labeled a troll. Nothing I brought up is unreasonable argument if you actually understand how to interpret data, experiments, or have a fundamental grasp on logical inference. You yourself acknowledge a .05% variance in the 4 gear sets' results which any reasonable person would accept as not being statistically significant enough to form the basis of the argument that determination is better than crit.

Anyways, I'm done responding to this thread since I'm a 'troll'. Official forums have lived up to their reputation of being devoid of intelligent discourse. Back to BG forums.

[Huntington]

Oh the internet. Where we are expected to respect random person X's data like it's just been published in Nature and is about to win the Nobel Prize in Chemistry without actually being able to question the results on the basis of the experiment or demand verification/replication of the results by a 3rd party. And if someone questions the self-proclaimed expert without their own equally spurious data they are automatically labeled a troll. Nothing I brought up is an unreasonable argument if you actually understand how to interpret data, experiments, or have a fundamental grasp on logical inference. You yourself acknowledge a .05% variance in the 4 gear sets' results which any reasonable person would accept as not being statistically significant enough to form the basis of the argument that determination is better than crit.

Anyways, I'm done responding to this thread since I'm a 'troll'. Official forums have lived up to their reputation of being devoid of intelligent discourse. Back to BG forums.

I think I said that same first bit myself to him in my terrible thread.

[T0rin]

It's fine that you should question my results, I'd hope that would happen. But if you are going to, you should actually read everything available to you before you say that nothing is verifiable. I've made everything I've ever concluded completely open and detailed specifically so people could replicate and verify or disprove it. Just because you didn't want to take the time, and then call me out on not doing what I've already done multiple times, DOES make you a troll.

I've had 100% accuracy with <400 accuracy on all extreme primals. The number you are posting only even matters for turn 5 snakes and even then it's too high post 2.1 since pet accuracy seems to be simply your accuracy and not the arbitrary 450 it was before.

Yes, you and everyone else can hit 100% accuracy on Ex Primals, nobody said otherwise. The numbers (450 for the pet) are for Turn 5, which is something everyone needs to gear for, unless they have their weapons and simply mean to abstain from the fight. However, for the majority of players who are still attempting or farming turn 5, the accuracy is very much a necessity. And no, post-2.1, pet accuracy did not revert to 435, try using something other than FFXIV-APP, and you'll find out, because it doesn't parse pet accuracy accurately, for quite a while now.

You aren't using food in your total # and if you are you should specify which you are using for those accuracy heavy fights.

I'm not using any +accuracy food, because accuracy doesn't affect the pet. Why am I being ridiculed by a GLD on how SMN mechanics work?

If the ridiculous BiS list you posted for both ACN and SMN wasn't enough this line is the icing on the cake of the bad advice you're putting forward.

You are also making some ridiculous gear choices, completely abandoning any crit for most of your slots. Crit is Summoner's #1 most important secondary stat. I happen to think there's a certain cut off point once you're in the ~550 range where certain gear choices towards DET start to make sense because crit gear up to 600ish max SMN can hit doesn't seem to produce too noticeable of a crit increase in parsers vs staying in the 540-550 range and getting more DET. But, you aren't even remotely close to that # with gear choices like allagan circlet or caster ring (or myth legs + circlet + tremor earrings, etc.)

CRT is not the most important stat from a point for point perspective. See this thread: http://forum.square-enix.com/ffxiv/t...Formula-Thread

It explains in detail, but basically, the damage formula was reverse engineered from in-game data sets (sit in front of a target for hours, spam a single skill with a known potency value and write down all the results, then derive the formula from multiple sets where gear changes, stats change, etc.) to come up with a specific formula: (WD*.2714745 + INT*.1006032 + (DTR-202)*.0241327 + WD*INT*.0036167 + WD*(DTR-202)*.0010800 - 1) * (Potency/100)

Go check it for yourself, but the formula is accurate across any gear or stats you want to throw at it, with any skill. It gives you a damage value (within the 5% acceptable range of RNG variation) that will predict what any gear using any skill will produce. From this, you can derive how much each stat (on a single point basis) contributes to damage.. i.e. if I increase DTR by 1 point, I get X damage. Or if I increase CRT by 1 point, I get Y damage. You can easily figure out (see http://forum.square-enix.com/ffxiv/t...=1#post1784123 for a method to how) how much each stat contributes to damage. The result of which is WD > INT > DTR > CRT > SS.

So no, CRT is not "Summoner's #1 most important secondary stat". The more DTR (or INT) you have, the better each point of CRT is, so it behooves you to take some DTR, at least. But, a single point of DTR will always be worth more than a single point of CRT, so there are certain pieces of gear (Allagan Circlet for instance) that will always be better than the myth alternative (Summoner's Horn), regardless of how much CRT you have, simply because of itemization. Some more CRT heavy gear (Summoner's Doublet for example) will be better than the DTR heavy alternative (Allagan Tunic) simply because it has better stat distribution. You don't decide the value of a single piece of gear compared to another on the stats of that single piece of gear, you decide based on the entire set that allows you to reach a specific level of ACC.

There is nothing magical about 550 CRT, it is an arbitrary stopping point. The rate at which you get crits never diminishes, so balancing CRT and DTR is just a matter of finding the right combination of gear to produce the most damage possible, of which the aforementioned set does. But, I will admit, the CRT heavy alternative produces only 0.5% less DPS than the DTR heavy set, which is good for flexibility, but if I wanted the absolutely utmost damage I could get, the DTR heavy set gives slightly more. But for reference, this CRT heavy set is only a very very small amount worse than the aforementioned set:

Allagan Grimoire of Casting
Summoner's Horn
Summoner's Doublet
Summoner's Ringbands
Allagan Rope Belt of Casting
Allagan Breeches of Casting
Allagan Boots of Casting
Hero's Necklace of Casting
Tremor Earring of Casting
Hero's Bracelet of Casting
Hero's Ring of Casting
Vortex Ring of Casting

SMN is more of a crit class than any job in the game besides Scholar. Bard has 2 DoTs who's crit rate matters for DPS increase (River of Blood trait and only 3 DoTs total) and even then it's only a 50% chance that the crit proc will trigger River of Blood ontop of certain fights like Atomos or Ifrit Extreme where 1/2 of the dots that matter for River of Blood can't be used.

SMN can crit the following every 3 seconds: Bio II, Miasma, Bio, Shadowflare, (this is already a higher # of things that can crit than BRD's 2 DoTs + Flaming Arrow every 30sec), Ruin I/II, Pet attack, Auto Attack, and Miasma II if its Contagioned. That's 5 attacks that you should have up 100% of the time and 2 more (Miasma II and Auto attacks) that get used situationally. So that's 5-7 attacks that all can crit independent of eachother. And that's not even counting the spell speed trait when your pet crits or Fester/ED criting. SMN is a crit based job through simple probability.

Now here is where you really start to lose it. BRD gets procs based on crits, procs that can actually influence DPS output by a noticeable amount. SMN can proc a spell speed buff, that while noticeable in terms of how much of a buff you get, occurs so infrequently (even with the most potential CRT you can have on gear) and lasts for such a short amount of time, that it comes out to be a 0.01% DPS loss using the CRT light set vs the CRT heavy set. It matters _that_ little.

It doesn't matter that you can crit 100 times a minute using 5+ abilities simultaneously. None of those crits can trigger the buff. Only your pet's attacks can trigger the buff, so you could be critting 10 million times a second on your DoTs and other spells, and you'd have a 0% increase in your chance to proc an SS buff. So I ask you, what exactly do all those abilities critting actually accomplish, other than dealing more damage? I'll tell you: Nothing.

In the end, CRT is just another way to increase damage output, just like DTR, just like INT, just like WD and just like SS. It is not special, or magical, it doesn't provide you any unique benefit. It's just more damage, and on a point for point basis, it isn't even the best source of damage.

Central flaw of the simulation he put forward that 'prove' DTR > CRT.

Actually no, the damage formula proved that DTR is better than CRT, my simulation was just a means to determin the impact of spell speed on an actual realistic rotation, and have some means to put some realism to all the math. There is nothing in the simulation that is biased towards DTR or CRT or any specific stat. That's the point, it is meant to be an unbiased test. If you can find any flaw in the simulation that would give out false data, please, let me know, and I'll gladly fix it.

- As he himself admits they are 'practically equal'. At the amount of damage he 'simulated' the few hundred damage difference between any of the 3 sets is not statistically significant. Nothing about his test was in any way conclusive in any of the directions he pointed to, especially since he 'simulated' it? The 'materials and methods' section of his paper requires a lot more information + providing the means for other interested parties to duplicate his findings. The baseline for statistical significance at an academic level is typically 1%-5%. Seeing as how he himself acknowledges all sets are within 0.5% of each other, his results don't really show anything that we can make definite conclusions about.

It isn't significant, and I've said that many many many many times. But, 'best is best'. If you want the absolutely utmost damage, there is only 1 set. It may only be very slightly (0.5%) more than other sets, but it is still 'best'.

And again, if you wish to duplicate any of my findings, I've documented in the above posts that I've already linked. See the damage formula thread if you want to see how they went about coming to it. See my documented step by step math if you want to see how I used that with SMN specific data to come up with some kind of analysis of stat benefits. Anyone can take that damage formula, go write down 1000 numbers in the game as they pop up on their screen, and come to the exact same conclusion that I and others have. It's all been documented, you apparently just haven't taken the time to read it. But, that doesn't seem to stop you from saying it is all wrong, with no proof of your own. And if you bothered to read it, you'd have known that RNG was taken completely out of the equation when it came to the simulation, it is all pure average damage (which reflects all the RNG in the game), applied to realistic spell rotations. That's why the simulations come out with the exact same damage every time you run them with a specific set of stats, because there is no RNG. That is why I can say definitively why one thing is better than another.

More Problems I see with his test
- All of the gear sets he tested rest on the assumption that 100% accuracy for Garuda is equal to maximized damage. He needs to prove this. If you're at 98% accuracy but you're doing 3% more damage per action then you are actually doing more damage than being at 100% accuracy with lower damage per hit. And since the difference between accuracy #s at the gear he's talking about vs gear with crit is at most 1-2 missed hits for the entire fight he needs examine other gear sets with less accuracy but more determination, crit, etc. Because don't forget you're already at a higher accuracy % than Garuda is and that crit/determination value applies to more than just your Garuda's auto attack.

I'll work on updating my simulation to allow you to specify and accuracy cap for 100% accuracy, and try to model out some miss chances and how they are affected by single and double digit accuracy deficiencies. There has been some 'sloppy math' done in this regard with real data that shows that the gap is wide enough to side with Garuda accuracy being more important than a small amount of secondary stats, but it is worth going further with to attempt to prove more definitively. But, these are still apples to apples comparisons, assuming 100% Garuda accuracy (which you might get against Ex Primals and non-BC content) for every set, and passing in stats of sets with almost 450 accuracy. It's not like I'm comparing a 448 ACC set to a 428 ACC set and saying "look, the 428 ACC set is so much better!"

- He's not counting food which again since this whole test is based around Turn 5 BiS he should taking account of since that is one of the few situations in game where you 100% should be using food and that changes the gear sets around significantly to hit his '448' number. If he is using food in the simulation he needs to actually tell us what food it is, because again the point of these kinds of tests outside of XIV is that other researchers can replicate it to verify your findings.

I am counting food, if you looked at the simulation logs (linked several times, including inside this post), you'll have seen that. And since food doesn't affect pet accuracy, and there is one distinctly superior food choice (buttons) to make, it doesn't change gear sets at all. But my simulation offers the choice to specify "food=eggs" if you really think that the extra CRT is superior, just to prove to yourself that it isn't.

Torin stated he considers Bard to be a crit based class and SMN not to be. My argument is that if SMN has more actions that can crit independently of each other every GCD / 3 seconds then Summoner should also be considered a crit based class. The more things that can independently crit per GCD, the more importance crit should logically have to you.

Why would that make it a 'crit based class'? You get no actual benefit from your ability to crit a bunch of times every 3 seconds. It literally has no actual impact other than another way to increase damage. Why does critting more times every tick 'logically' mean it is important? DTR affects every single attack all the time, does that mean it is superior? Well, not based on just that logic, no. Doesn't make any sense at all.

[T0rin]

So far his arguments have been more compelling than yours, and he clearly has a firm grasp of what he's arguing. Labeling him a troll is the most severe cop out possible, but you're likely not capable of producing a better argument than he, so I guess your "troll" response was expected.

Clearly he has a firm grasp on what he is arguing, even if he is obviously wrong about how SMN mechanics work? Really? I'm guessing you're just his buddy from Hyperion who came to comment about how he was butthurt about being called a troll, which he is. I labeled him a troll in my post how I would actually write out proof of how he is wrong. It wasn't a cop out, it was me tagging the thread so I'd know where I'd need to come back to in order to dispel some dubious bad-SMN-logic that was being spewed.

Yep, I was right, same FC even. Please go back under your bridge. You really add a lot of credibility to his trolling. *cough*

[Pwnznewbz]

T0rin - /golfclap. Well said.

[AhmeraMae]

It explains in detail, but basically, the damage formula was reverse engineered from in-game data sets (sit in front of a target for hours, spam a single skill with a known potency value and write down all the results, then derive the formula from multiple sets where gear changes, stats change, etc.) to come up with a specific formula: (WD*.2714745 + INT*.1006032 + (DTR-202)*.0241327 + WD*INT*.0036167 + WD*(DTR-202)*.0010800 - 1) * (Potency/100)

Valks formula is an approximation and known to prdouce incorrect results. Also, valks formula does NOT apply to each and every job and skill in the game.

Edit: Another theorycrafting source, working on the actual game formula.
http://www.chocobro.com/forum/viewtopic.php?f=6&t=98

[T0rin]

Valks formula is an approximation and known to prdouce incorrect results. Also, valks formula does NOT apply to each and every job and skill in the game.

Edit: Another theorycrafting source, working on the actual game formula.
http://www.chocobro.com/forum/viewtopic.php?f=6&t=98

It isn't Valk's formula. Or maybe Valk adopted it after EasymodeX did significant work on it. Either way, this particular formula does apply to each and every job and skill in the game, that's the point.

And Eein's work (that you quoted) is 6 months old at this point. He's admitted to needing to revisit it, but AFAIK, has not published anything new. Regardless, it is easy enough for anyone, with any class, to verify the aforementioned formula.

[Zorzi]

The only reason I come to these forums nowadays is to watch T0rin lay smack downs.