King's Justice vs Raging Rush

[FailureMidgard]

While mathematical formulation is the way to go to see which CAN be better, parses are far more realistic when arriving at the conclusion as to which one IS better. It's irrelevant which one wins on paper with no buffs, only on colibri, when the moon is full, at night, if you engage every mob on an odd integer of vana'diel time. What's relevant is how they stack up in actual situations with actual buffs and actual human error/randomness thrown into it. You can't just say "Oh well shit, only 1/5 of the mobs were dia'd!" and just rule the parse as invalid because in any merit situation, that's what's going to happen. You can't say that MMJSP invalidates the parse because it's a merit camp that people actually go to. I really hope I didn't just somehow agree with Starr.

You can isolate variables though. If one benefits more from mobs being diaed, you can figure that out separately, then when you get to a party, try to watch how often the mage dias. If he dias often, use X, if not, use Y. Don't just look at one parse once and base everything you do off that.

Math is good, but it's biggest use is knowing what affects what how. Past that, for many gear decisions, parses are needed as a baseline to do the math with. My WS/TP splits aren't done with theoretical values of average melee hit, average weapon skill, 5 swings per WS with a GK, etc. They are based off many parses. I know how aggressively I usually play, and that varies from person to person. For all the talk about an X-hit build, the vast majority of players still get distracted, and many times end up WSing at X+1 or X+2, or end up with more TP because a mob was at 10% when they hit 100TP, etc.

And yes, you can rule out any single parse at MJSP that is trying to conclusively prove something. Get 30 different parses, and perhaps you have a point, but linking 1 MJSP parse as proof that X is better than Y is fucking garbage. I can conclusively prove that Subtle Blow increases my WS dmg with a MJSP parse, if you really want (compare a parse of /drg and /nin and see which has higher WS #s).

Parses are good to use, people should know where their parse results will vary from the theoretical formulas, but anyone who tries to use them as end-all-be-all proof is wrong.

[Starr]

If you really want everybody to come to a conclusion you need to be fighting colibri without exp buffs(cor roll), atk/acc buffs, or dia, ever. And then show averages on the two different levels of colibri.

That really wouldn't show anything though, I don't merit without support and I normally hope for Dia.

What's relevant is how they stack up in actual situations with actual buffs and actual human error/randomness thrown into it.

This is what I'm trying to say but of course human error and randomness are never constant so no parse is ever going to be comparable to another because of this but then on the other hand no paper math can be 100% accurate because it doesn't consider these things. I do however think that they should be more valid than just paper math alone, I don't think anyone got lazy near the end of that parse, it was so short that that wouldn't happen anyway. The mobs were NEVER Diad so it's not like there could be inconsistency there, the only inconsistency could be if Minuet was put up in time and Chaos Roll and the number of it which I do admit could have massive dispcrepancy, but even in a no buff situation DA could fuck up any results because it's not a constant, I did my best to keep the parse as fair as possible and even made it biased towards KJ with buffs and I nowhere said it's the be all end all. I could parse every party I ever join from now but there's no point cus people will find the stupidest, smallest, irrelevant reasons to dismiss.

[Knyghtmayre]

Get this straight, I'm not saying that the one parse shows which is better, I'm saying that as a total collection of data, it's all the same mobs, and if collectively 30 parses of 150+ mobs each are put together, a (semi-)accurate portrayal of which WS that person should use in that scenario should be able to be extracted from that. Dia on 1/5 of the mobs in one parse out of 30 parses is .06% of the total data set. The milliseconds you can shave off of how fast you engage/switch to each mob will swing your parse by more than that. Of course by that time it's likely that the person doing the testing won't need to know which is better in merits because they'll be capped on merits. Basically, there is no point in parsing ideal controlled conditions whenever conditions are never going to be that controlled in actuality. Sure it would be nice to think that your RDM will Dia every mob, or that you won't D/C on Chain 56 and lose all your buffs for 3mins upon returning, but it's not always going to happen. You should parse neutrality obviously, but who's to say that PnL didn't? I really doubt that he was trying to influence one WS's DMG over the other, as it's still a parse and trying to do less damage WSs just goes against the entire ideal of why he's parsing in the first place.

[Starr]

I know you're not, no parse will.
I have no ideal how PnL can parse WS that high with Just 2 Minuet and Hedgehog Pie with that gear as logically: FFXIGEAR.com his gear with Aggressor down is really poor on Accuracy, I don't know how he does it.

[PeaceNLove]

No. i didnot look at them as soon as i saw :

- parses posted for ALL mobs...
- sample size : 175 mobs with 40-56 puks, which means for each mamool jobs/level, skoff etc... a sample size less than 10 mobs...
- Dia used erratically, like poster said : 1/5 of sample size is out...
- MMJSP is an horrible camp to use parses for tests, because of debuffs a player can receive.
- chaos roll used on 2nd parse...

all these reasons make i can't even hope to calculate even a slight close avg value. In other words, those parses are useless fortopic.

On good side tho, poster posted at least his gears and used BRDx2 in 1st parse, and debuffs his rdm used... good start but way not enough sadly.

i'll quote Thunder "Only way I see an accurate parse is when you're parsing an instance where you're ONLY trying to prove the difference, not when you're getting merits,..."
If poeple don't get that (especially Starr), then can close this topic...

The second parse was just there to illustrate typical results I see every time I party at Jade. It wasn't a face-off between KJ and RR, just a demonstration of the type of numbers I get when mainly using KJ and saving RR for select occasions (Bergressor down, Lurkers, etc.) as the number used of each WS clearly shows (81 KJ vs 45 RR). The first parse was my "fair" fight between the two, but a number of people bring up good reasons to invalidate it, which is fine. Think of it what you will. Sadly I rarely merit at birds, and if I do it's on SAM so I can't provide any parses on those.

I also posted these parses in response to Starr's parse, where he only averaged ~920 KJ with an abysmal high of 1356 with similar buffs (though I'm not sure if you were eating yellow curry buns there or not?) on a group of mobs with equal or lower defense (on average). Not implying he's doing it wrong, but the fact that my average KJ nearly rivals his high seems questionable.

I have no ideal how PnL can parse WS that high with Just 2 Minuet and Hedgehog Pie with that gear as logically: FFXIGEAR.com his gear with Aggressor down is really poor on Accuracy, I don't know how he does it.

It is low on accuracy, but it's tailored to how I alternate KJ and RR in merits. Generally though, I've had no accuracy problems with it on anything short of Infiltrators/Lurkers regardless of Aggressor (unless I'm missing some potential DAs and don't know it). Can always sub in Champion's Galea now too for extra acc.

[Yugl]

Starr's parse needs more data in order to prove a point since it doesn't account for the fact that grips change DoT DMG. I didn't check PnL's data extensively, so I cannot offer criticisms.

No one is saying that parses are bad, but they absolutely are not the means of reaching a WS1>WS2. There's as much of a chance skewing the parse the second round as much as there is the first time. Parsing over time may even out skews or it may skew them even further.

For example, lets say I join a PT and Angon is always on when WS1 goes off, but is only on (8/10) times when WS2 goes off. Obviously the parsed is skewed in favor of WS1. This may not be a problem in Starr's case (And I think he even said Dia was not used?), but what about the COR's rolls? He may roll an eight (For Chaos roll this is bad) when WS1 goes off, but an 11 when WS2 goes off. Parsing overtime and multiple times does not change this since there's just as much chance that this scenario repeats itself in future parses. Would this be fair? Does using WS1 mean your COR is going to get an eight? Not really, it just happened to get bad rolls when you parsed.

It's not about obtaining "Ideal" conditions, it's about fair conditions. Another way to look at it is: If I replaced those times I did WS1 with WS2, would the result be different? If there's even a slight possibility of that occurring (Especially in this case where they seem close), then you cannot state WS1 > WS2 (Even through 10000000 parses). Math bypasses this by placing them in *fair* scenarios. This is increasingly so for WSs since "ideal" conditions are less difficult to achieve (Note: In cases where you'll save TP for the next mob, it does skew the parse since that excess TP could have gone to the OTHER WS as extra DMG).

Once more: Especially for WSs, math places them on *fair* conditions, not simply *ideal* situations. The best part however, is that this is easily controlled in game if you want to waste A LOT of time. You would basically need to get the COR to roll the same number (Aka bring him and you to the point of wanting to kill yourself), have both BRD songs on, no dia (So you can sleep the SAME bird in between WSs), and all other buffs the exact same. You can do this if you want, but why bother when you can ask people to pen and paper it for you?

Edit: Another example that might be clearer. We're all familiar with the 400 ACC on birds = cap (95%) right? Well lets say I went with a Thalassocrat and happened to get the EVAs proc on my 380 ACC set during points where otherwise, I would have missed. Does this means I'm at 95% ACC (Which would show up on parses) and that it's equal to my 400 ACC set? Nope. It just happened to be at that ACC due to pure luck. There is ONE point I'll give to the parsing Tarus though (And even I'm not sure if this point is correct). Aren't formulas based on parses? If so, that's a skew right there.

[Starr]

I also posted these parses in response to Starr's parse, where he only averaged ~920 KJ with an abysmal high of 1356 with similar buffs (though I'm not sure if you were eating yellow curry buns there or not?) on a group of mobs with equal or lower defense (on average). Not implying he's doing it wrong, but the fact that my average KJ nearly rivals his high seems questionable.

Was with Yellow Curry (attack+78 STR+5) and Chaos Roll on top of yours. I would like to see more of that parse to see how you do it, I swear you just don't WS on Lurkers/Infiltrators or with Berserk/Aggressor down or something.

[Thunder]

This is what I'm trying to say but of course human error and randomness are never constant so no parse is ever going to be comparable to another because of this but then on the other hand no paper math can be 100% accurate because it doesn't consider these things.

But if one wins with math it will over the course of your career as war be better because, unless a missed buff will put one ws over the other because the misses in buffs/debuffs cannot be foreseen AND FOR THAT REASON a parse cannot be accurate because you derive your conclusion from the inaccurate information and erratic buff/debuff frequencies of a party. But with math you APPLY conclusions to the erratic buff/debuff frequencies.

So why has nobody mentioned masa's math? I looked through it again and it points to RR > KJ always. Nobody disagreed with it, either.. so is that the general consensus of the math community?

[Starr]

I don't know, really though at 100% TP with KJ fTP=1.0 the only difference between the WS' is that 1 STR on RR=0.35*0.83=+0.29 weapon damage and 1 STR on KJ=0.5*0.83=+0.415 weapon damage so for every 10 STR, and there are only a few slots where you can change this from RR cus most STR gear has DEX on it, you would gain 12.5 weapon damage which multiplied by pDIF of about 1.5 and by 3 for number of hits is only +31.5 total damage ignoring RR criting at all which will add at least 100 damage per crit. Add in the fact that you are going to be wearing more DEX for RR therefor having more accuracy and the fact that RR can and often does critical hit and I don't see how KJ can be so awesome.

[PeaceNLove]

Was with Yellow Curry (attack+78 STR+5) and Chaos Roll on top of yours. I would like to see more of that parse to see how you do it, I swear you just don't WS on Lurkers/Infiltrators or with Berserk/Aggressor down or something.

Well, my set that you called poor on accuracy actually has 4 more than your KJ set when you factor in gorget acc. Those statements are pretty much true, though, for the second parse. In the first parse there were no Lurkers, but yeah generally I RR when I realize my KJ is gonna suck lol (doesn't apply to the first parse). The pulling BRDs I merit with usually sing a madrigal right before Lurkers as well if time allows.

[Cream Soda]

The pulling BRDs I merit with usually sing a madrigal right before Lurkers as well if time allows.

Yuck

[Starr]

Well, my set that you called poor on accuracy actually has 4 more than your KJ set when you factor in gorget acc. Those statements are pretty much true, though, for the second parse. In the first parse there were no Lurkers, but yeah generally I RR when I realize my KJ is gonna suck lol (doesn't apply to the first parse). The pulling BRDs I merit with usually sing a madrigal right before Lurkers as well if time allows.

u mad? That is my bird set, for Mamools I use Pizza+1 which is liek +35 Accuracy.

[PeaceNLove]

No, what made you think I'm mad? I enjoy discussing WAR stuff. I also find pizza to be, on the whole, really unnecessary. I parse 89-92% accuracy at MMJSP regardless of food. I've had another WAR with identical accuracy gear/merits minus Ebody eating Coeurl Subs parse 0.7% accuracy below mine while I ate Hedgehog Pie. But then again I party with BRDs that madrigal for Lurkers... What's so "Yuck" about that, anyways?

[blackmail]

But if one wins with math it will over the course of your career as war be better because, unless a missed buff will put one ws over the other because the misses in buffs/debuffs cannot be foreseen AND FOR THAT REASON a parse cannot be accurate because you derive your conclusion from the inaccurate information and erratic buff/debuff frequencies of a party.

Words are important too. Don't neglect them.
Because Synonym | Synonym of Because and Antonym of Because at Thesaurus.com

[Sifer]

So why has nobody mentioned masa's math? I looked through it again and it points to RR > KJ always. Nobody disagreed with it, either.. so is that the general consensus of the math community?

The genral consensus of the math community is that no one wants to take part in this thread if they get flamed for doing a simple calculation. If you're so sure about which WS is best then you shouldnt need this thread.

Starr= the new Ilean. everyone is obviously retarded for using haste gear.

[Starr]

I haven't flamed anyone, and I am a bad man for pointing out errors that make certain people's math pointless now? Talk shit all you want, I will still be the better Warrior Sifer.

[Sifer]

lol

[Starr]

Yeah you are pretty funny.

[mallister]

I party with BRDs that madrigal for Lurkers... What's so "Yuck" about that, anyways?

because as a 2h melee you can eat pizza +1 and keep chaos/double march or whatever other buffs float your boat without an accuracy dip?

[Yugl]

because as a 2h melee you can eat pizza +1 and keep chaos/double march or whatever other buffs float your boat without an accuracy dip?

If I remember correctly, it was about 500ish ACC to cap on Lurkers.