Maximizing SA/TA Damage (mathy mathy)Follow

Some useful LJ links from when I asked about the worth of heca for solo JAs:
http://releenaseraph.livejournal.com/112302.html
http://releenaseraph.livejournal.com/114592.html

Edit:
Quote of importance:


Edited, Aug 9th 2009 1:26am by Noodles

Thought so, I knew I read this somewhere but I couldn't find the source.

Now, let's do a math redo!

Main: Blau Dolch [DMG:33 Delay:178 Acc+5 Attack+16]
Sub: Sirocco's Kukri [DMG:26 Delay:150]
Range: -
Ammo: Fire Bomblet [Acc+6 Attack+6]
Head: Anwig Salade [Haste+3% Acc+10 Attack+5]
Neck: Love Torque [Dex+5 Dagger+7]
Ear1: Brutal Earring
Ear2: Suppanomimi [DW+5%]
Body: Homam Corazza [Acc+15]
Hands: Homam Manopolas [Acc+4 Haste+3]
Ring1: Rajas Ring [str+5 dex+5]
Ring2: Toredor's Ring [Acc+7]
Back: Cuchulain's Mantle [str+4 dex+4 acc+5]
Waist: Swift Belt [acc+3 attack-5 haste+4]
Legs: Skadi's Chausses [acc+4 attack+5 haste+2]
Feet: Homam Gambieras [acc+6 haste+3]

Level 82: Defense 327, Evasion 339, VIT 67, AGI 67
Coeurl Sub: Str+5 Agi+1 Attack+20(75)
1BRD: Attack+66 Haste+11%
haste: Haste + 15%

str: 62 + 9 + 5 [76]
dex: 80 + 14 [94]
attack: 8 + 292 + 38 + 27 + 66 + 75 [506]
Acc: 200 + 82 + 47 + 65 [394]
Haste: 15 + 11 + 15 [41]

fstr = 3
cratio = 1.197
pDIF:[Min = 0.936][Max = 1.436][Average = 1.186]
Acc% = 88.5
Base Damage = (33 + 3) && (26+3)

Attack Round = ((36 * 1.186) + (29 * 1.186)) * .885
= (42.696 + 34.394) * .885 = [68.22465]

Delay: (178+150) * .8 = 262.4; 262.4 * .59 = [154.816]

--------------------------------------------------

Main: Blau Dolch [DMG:33 Delay:178 Acc+5 Attack+16]
Sub: Sirocco's Kukri [DMG:26 Delay:150]
Range: -
Ammo: Bomb Core [Attack+12]
Head: Hecatomb Cap [Str+11 dex+5 Slow+9]
Neck: Love Torque [dex+5 dagger+7]
Ear1: Brutal Earring
Ear2: Pixie Earring [dex+3]
Body: Skadi's Cuirie [dex+8 attack+5]
Hands: Hecatomb Mittens [str+7 dex+4 slow+5]
Ring1: Rajas Ring [str+5 dex+5]
Ring2: Thunder Ring [dex+5]
Back: Cuchulain's Mantle [str+4 dex+4]
Waist: Warwolf Belt [str+5 dex+5]
Legs: Hecatomb Subligar [dex+8 attack+20 slow+12]
Feet: Hecatomb Leggings [str+6 dex+3 slow+4]

Level 82: Defense 327, Evasion 339, VIT 67, AGI 67
Coeurl Sub: Str+5 Agi+1 Attack+20(75)
1BRD: Attack+66 Haste+11%

Str: 62 + 38 + 5 [105]
Dex: 80 + 55 [135]
Attack: 8 + 292 + 52 + 53 + 66 + 75 [546]
Slow: 9 + 5 + 12 + 4 [30]
Haste: 0 + 15 + 11 [26]

fstr = 10
cratio = 1.319
pDIF1:[Min = 2.000][Max = 2.582][Average = 2.291]
pDIF2:[Min = 1.000][Max = 1.582][Average = 1.291]
Base Damage = (33 + 11 + 135) && (26 + 11)

Attack Round = (179 * 2.291) + (37 * 1.291) = (410.089 + 47.767)
= [457.856]

Delay: (328 * .85) = 278.8; 278.8 * .74 = 206.312;
206.312 * 1.3 = [268.2056]
-------------------------------------------------------

Main: Blau Dolch [DMG:33 Delay:178 Acc+5 Attack+16]
Sub: Sirocco's Kukri [DMG:26 Delay:150]
Range: -
Ammo: Bomb Core [Attack+12]
Head: Skadi's Visor [Dex+4 Attack+6 Haste+3]
Neck: Love Torque [dex+5 dagger+7]
Ear1: Brutal Earring
Ear2: Pixie Earring [dex+3]
Body: Skadi's Cuirie [dex+8 attack+5]
Hands: Enkidu's Mittens [str+4 dex+4]
Ring1: Rajas Ring [str+5 dex+5]
Ring2: Thunder Ring [dex+5]
Back: Cuchulain's Mantle [str+4 dex+4]
Waist: Warwolf Belt [str+5 dex+5]
Legs: Hecatomb Subligar [dex+8 attack+20 slow+12]
Feet: Enkidu's [Dex+3 Attack+4 Haste+2]

Level 82: Defense 327, Evasion 339, VIT 67, AGI 67
Coeurl Sub: Str+5 Agi+1 Attack+20(75)
1BRD: Attack+66 Haste+11%

Str: 62 + 18 + 5 [85]
Dex: 80 + 54 [134]
Attack: 2 + 292 + 42 + 63 + 66 + 75 [546]
Slow: 12 [12]
Haste: 5 + 15 + 11 [31]

fstr = 5
cratio = 1.319
pDIF1:[Min = 2.000][Max = 2.582][Average = 2.291]
pDIF2:[Min = 1.000][Max = 1.582][Average = 1.291]
Base Damage = (33 + 5 + 134) && (26 +5)

Attack Round = (172 * 2.291) + (31 * 1.291) = (394.052 + 40.021)
= [434.073]

Delay: (328 * .85) = 278.8; 278.8 * .69 = 192.372;
192.372 * 1.12 = [215.45664]

----------------------------------------------------------

Damage Difference from SA: (457.856 - 434.073) = [23.783]

Delay Difference: 52.74896

Damage Difference from Attack Round: (52.74896 / 154.816) = 0.340;
(.340 * 68.22465) = [23.196381]

The Hecatomb does better when it's supposed to be worse. I don't believe I've done anything wrong, but I do want somebody to check over it and make sure I didn't screw anything up. The first thing that made me say "wait.. I don't trust this" about the list is "skadi = DH nq". 5 attack is not as good as 2dex for sa.

edit: Found a small pdif error, fixed it; numbers are closer by about .5 now.

Edited, Aug 9th 2009 12:24pm by Deadgye

Would the second march make a differance against the hecatomb? See quoted section in last post.

I'd like to see the derivation of this result.

EDIT: heh, Banalaty and I are of one mind on this. I can come up with a few good ways to test it (Ceremonial Dagger + full Heca on Uragnites, anyone? :-D) but it still sounds like about as much BS as the current "this is what's being said about the 2h cRatio -> pDIF formulas on BG" is.

Edited, Aug 9th 2009 9:00am by MDenham

Actually, im pretty sure DA/TA still put more weight on melee rounds. A single TA will double your melee round damage (4 hits instead of 2 with 50/50 split of it being your main or off hand so averages out over time past the different base damages). A TA will not double your SA damage, just add a static 50% of a melee round damage to your SA round.

10% TA and 5% brutal will give you a straight melee increase of (i think) 24% on top of everything. Even with the 100% acc on SA/TA, the increase in damage from multi hits is obviously going to be far less than 24% since it does nothing to the huge crit hit. Basically, multi hits add far less of a % increase to SA/TA than it does to a melee round. So adding multi hits will increase the damage of your melee rounds by 24% while only adding say 5%(made up number) to SA/TA power.

The thing is, no matter what gear you use on SA/TA you will still get your extra multi hits for SA/TA rounds regardless of your slow% or even if you didnt use SA/TA at all. It is basically a static bonus on every attack round. But by slowing down the following attack round, instead of slowing down 50 damage, your really slowing down about 62.5 damage. So if you have enough slow to cut your delay in half (using whatever slow calculation is right, still loosing half the delay), you would be cutting off 31.25 damage from slow, not 25. The fact that you still get the ~12.5 damage from multi hit added to a SA/TA round is irrelivant because you would get it even if you didnt use SA. Your gear choice wont change it (well slight bump to Fstr, but ultimately the damage change wont be significant).

Changing heca hands to enkidu hands will have almost no effect on the multi hit damage on the SA/TA round. You will still get your damage from multi hits no matter how much slow you have. But, it WILL have an effect on the following rounds total damage which DOES include multi hits. Upping the damage of your melee by 24% means the detrimental impact of slow is increased by 24%, while having virtually no impact on the SA/TA round because multi hits on SA/TA rounds are independant of using SA/TA.

Does that make any sense? Or am i completely off base?

Alright guys. So it's seriously my bad for posting this and never responding to it. Big apologies. And sorry for bumping this as well. I was on vacation with no computer for much of last month, and it kind of made me forget to come back to this.

I'm going to respond to a couple of concerns some of you had, though obviously many of you knew enough to figure some of these things out yourselves. However first, I would like to say that I've edited my LJ article on this (my main area for information-keeping) because of the critical miscalculation that some of you discovered. That is, the function of slow%. That is, I'd assumed based on outdated information that gear slow was a separate function, and not the -haste% that it actually is.

Edits
It's a tribute to how annoyingly complex the project became, that I spent many hours last night trying to amend this. And I had to re-learn how I put the formula/spreadsheets together. The most annoying part was that moving from office 2000 -> open office -> office 2007 caused some annoying corruptions.

But I fixed it,
*I changed the function for delay in the formula
*I recalculated ~1000 sets of equip, and found the new "best set"
*I used that for the new template for cross-comparing stats, and thus reworked the ratios

Notes:
*I only had time to do SA last night, not TA
*The gear rankings are now slightly out of order and need updating
*New spreadsheets have yet to be uploaded (the old ones were probably partially corrupted anyway)
*More editing yet to be done

I've already updated the OP. Mainly all that changed was that slow hurts more (obviously). Will update again when the rest is finished.

Misc Concerns
Now to concerns,
Sorry about that, and thank you for the suggestion. Out of pure forgetfulness, I neglected several pieces of gear, and they've now been edited in (on LJ).
No. More on this shortly.Same, more on this. And yes, you can edit the spreadsheet. However, it would be very complicated. It sucks for me to make edits like that, and I designed it -.-. I'd recommend taking a look at least to see if you can figure it out. What you'll want to do is look at the area where it says "formulae composites." There isn't a single cell with the formula. Try and look at how these cells are working. As a note, I use pDif avg, (Min+Max)/2. After editing the formulae composites in a single row, you'll have to copypaste for all rows in the data table.Almost everything has been taken into account. Including fStr and atk from str. I made the formula to be very accurate, and I would trust it above any alternative less complex. The largest factor unaccounted for is the TP-gain differences, and their relative associated WS dmg contribution. Thus, the importance of d.a./t.a./haste are slightly higher than shown, and slow lower.

Using this as a Guide

I understand what you mean, but it actually it does hold water so to speak. I started this project, like many others, to maximize my own utility as opposed to anyone elses. Then I wanted to pass on what I'd found. But I realized that without creating stat ratios, I wouldn't be offering anything too tangible. So I developed those as a guide. And that's precisely why they are very effective to have. Because now we have a base to refer to, whereas before, we didn't. In my example, ~2.2atk = 1dex for sa. Use this as reference. You know the value of dex will be lower if you have more dex, for example. It's impossible to make a "perfect" guide for rationalizing stats, but it's better to have one than not. I/we could make a much better spreadsheet or calculating program for plugging in your own stats, which would be best, but it is too complicated and time-consuming for me to do.

To reemphasize, please use this as a convenient guide. As the math veterans know, it is not be-all/end-all.

1. On Differences in haste
I probably shouldn't have done this in double march from the beginning, but I wanted to make a good case for haste pieces, since to many they were still the underdog. Fortunately it wasn't too difficult to add numbers for single march+haste and solo haste spell either. Keep this all in mind when looking at gear comparisons; think relatively, or calculate using the other available haste ratio values.
2. On Differences in attack/pDif
This is obviously a problem. However it's not terribly bad. On HNM's we tend to sub war and use pizza. On colibris, /nin and crab sushi (at least I do, but I know everyone's different). We also might get buffed. Dia will be more consistent. This will compensate a little. I usually do chaos roll in exp pt's, and I don't mind that my #'s don't account for it. As mentioned above, stray from the guide and use it as a starting point. One more thing,

The only times that the settings herein (being specific as they are) will lead to inaccuracy is when the comparison between two (sets of) gear pieces come out to be extremely close. Here's what I mean,

Not Close
Single March, Haste
Skadi's Visor, 24.12
Hecatomb +1, 3.48
(when the numbers come out like this, who cares about being 100% accurate for every situation?)

Close
Single March, Haste
Skadi's Visor, 24.12
Maat's Cap, 26.46
(depending on the situation, this could go either way)

What I'm trying to show with these two examples is that the rankings and stat ratios can be very convenient and even quite accurate, even as rough guides. This may not be so helpful to the forum vets with veins of which internet arithmetic doth flow through, but for others it might be more helpful.

It's akin to parsing, which shouldn't be taken to give you irrefutable conclusive answers, but yet can be very helpful for zeroing in on desired data, and still be quite accurate in deciphering solid from utterly bogus.

Don't worry about that. I calculated everything, then found how much a single point more of each stat would produce. The stat to pick to compare the other stats to is arbitrary, and you'll get the same relative ratios regardless. I picked d.a. because it was the lowest number.

Edited, Sep 24th 2009 11:56am by Shamaya

Wall of texts aside,
Trick Attack Data
I've now edited this on top of the previous SA edits. So now everything should be up to date. What I changed:

*changed slow% to be -haste%
*added 1minuet's worth of attack to make up w/ the inconsistency since the attack values in the spreadsheet were all riding the area between two tiers, skewing the comparisons. attack now has a higher value, and is now in the same tier of attack that the SA set is in. keep this in mind if possible
*edited dex to be quantifiable for TA

This project is definitely not perfect. While the formula used is almost perfect, the application of it to come up with these stat/gear ratios/listings is certainly relative. I caution again to use this as a guide; while it is a decent one it is a guide nonetheless.

I was actually just editing that that as you posted. That list had not been updated to the new ratios. It is now updated though, and also edited into this OP. Your suggestion has also been added, thx. Really sleepy now

I like you removed "movement speed" as a variable in your results, although I disagree with "Strider Boots / Skadi's Jambeaux are, essentially, the best Sa/Ta feet you can wear at almost all times."

Even after your post of clarification, I guess we'll just have to disagree and go about our buisness.

But considering the Delay from popping the ability, and the fact that you can lock on to get behind the mob/player, I don't see it as a valid point at all.


Even if you dont fight locked on, you can still lock on to circle, then unlock. The only time you may not want to circle around a mob like that is very large mob (like a HNM) or something with an AoE you want to stay clear of (Such as a Chariot boss). In cases like this all this about maximising sa/ta through haste/slow is worthless anyway.

In fact I'd like to add a point to make sure that all of you Thfs who are blindly changing your SA/TA sets because of all these considerations, are still keeping a 2nd SA/TA set, using full power and ignoring haste/slow, to use with harder mobs that you don't TP burn like a merit mob. Things that you maye run in to SA/TA/WS and gtfo again, or harder mobs without a cor or brd. Like a sky god without melee buffs where your melee damage will be so poor anyway that the majority of your damage comes from sa/ta.

Edit: Also, you missed Oily Trousers in all cases, which would beat the dex/agi+4 onry options, as they have dex/agi+5.

Edited, Sep 26th 2009 1:30am by Noodles

Yes you do. It's not the animation that causes the pause, it's the JA it's self. It doesn't matter whether you blink or not.

Edited, Sep 26th 2009 6:18pm by Noodles

Nah, my post was that you shouldn't ever have trouble running to position yourself that dusk/striders will make any difference, on smallish mobs (like merit mobs), considering it's quicker to strafe behind them whilst locked on, and you will still get some JA-delay to give you an extra 0.5 second or such to get inline. It shouldn't be a problem.

Alternatively you could, you know, move into position when your timer is at 1sec, making all movement and positioning related points irrelevant.

And don't try and argue that "the mob might move in that split-second between moving and using the JA!". Because sure, it's true. But the mob could also not move at all untill you've used the JA then moved into position, but still move just befor you hit, caussing you to miss anyway. It can theoretically move anytime, it's just bad luck, but it's a constant that can be applied to every situation, and not worth using as an argument for or against XYZ.

Edit: And as another post, if your countdown is approaching 0, is it really too hard to wait untill your next attack round befor popping the JA? so you've saved 0.5secs on that JA, great. Chances are it wont matter because of time wasted running to/engaging/whatever the next mob. But with that half a second you've "wasted", waiting for your next attack round, you now have ample time to position yourself befor the next attack round.

I'm banging on, but I really see no argument whatsoever for/against movement speed +/-.

Edited, Sep 27th 2009 12:33am by Noodles

Bagh, no use in going into this further really >,<. If we can't see eye to eye by now then it's not worth trying. Because no, strafing is irrelevant to me since it's ineffective given how fast we (can) attack, and depending on the degrees turned. It really isn't difficult to unlock and is more reliable, I'm not sure why anyone would strafe unless it didn't turn that many degrees (also, I'm talking about TA here, not just SA, players also move). As for moving into position when the ja comes up..who wouldn't do that? You also said 'don't try to argue they could move at the last second'--that's precisely what I'm arguing, otherwise this would be about irrelevant. No, it's no gaurantee you can save a lost sa/ta every time. No, the mob/player won't move every time. Hence why I said it turnes out to be %probability*hit(movement) vs 100%probability*(hit(+dmg feet) - hit). And yes, timing every sa/ta is not worth it, being tedious. And actually it wouldn't make a difference, as if you hit the ja when you have full delay, you have a longer window for something to happen, and if you hit it right before you are about to attack, you won't have enough time to respond. Changing the feet isn't about avoiding missed sa/ta's (before you unlock), it's about speeding things up for when you do try to avoid missed sa/ta's. Oftentimes, unlocking will be more effective--period. Please don't forget that this is also only one of the points I mentioned. If you want though, you could also try to only use movement feet for sa/ta during certain fights; it is cumbersome, but certainly doable. I'm not sure that someone who wouldn't bother to change the feet normally would bother changing them for special fights (like chariots) tho. Forgive this WoT:(