STR Rings V. Acc Rings (Yes I've searched already)

[aurik]

All of the time, its better to setup for optimal group performance.

I find it hilarious that if a Flame Ring increases your damage by say, 2-3%, it is clearly useless and totally worth insulting people over. But if a Sniper's Ring +1 increases your accuracy by 4-5% (relative increase, if your acc was 70%, +3.5% acc is a relative 5% boost) it is clearly a world shattering difference, and will make you a damage god. Of course an accuracy ring will make a greater difference than a Flame Ring when using attack food and accuracy is far from cap, because it is improving the weakest point in the damage calculation. That doesn't mean the STR actually does nothing even then, unless you're fighting a mob type with very low VIT, it just does slightly less.

In reality, sometimes sushi is just more practical, for example with a mixed party where some people are using lower rank/unmeritted weapons who would still struggle with accuracy with as much as they can equip on, or against a selection of mobs that includes a significant number of THF mobs. As a MNK, accuracy even plays an additionnal role because an accuracy check is required for potential counters to activate, so greater accuracy for a MNK results in less damage taken and more damage caused when tanking without shadows/between shadows. A good party with good mob availability at colibris can get 20-25k avg 30k/hour xp peaks with attack food or with accuracy food. Is it really that important if they're eating fish or red meat?

In most situations, you don't need the full amount of accuracy sushi gives, making it somewhat wasteful. It doesn't mean its always useless, and using attack food doesn't make you a better person, it just increases your attack. What makes sense is being ready to adjust and make the best of it either way.

Except this discussion isn't about "how I get 20k in Colibri"; that can be done nowadays with 4 blind puppetmasters and a meleeing red mage. This is about "what is better for my character in the ring slot, strength or accuracy".

When one type outperforms the other by a relative 50-100%, it isn't one type of ring edging the other out, it's a fucking rout.

[Nagamaki]

Here's a situation with some made up numbers :

NIN
Senj/Perdu so D38/D32
Ring options for 1 finger : Flame Ring or Sniper's Ring
fSTR against imaginary mob's VIT : 8
Mob defense : 350
Average pdif : 1.6
Average effective accuracy : 75%

These are all seem like fairly reasonable numbers. I have NIN levelled but I never use it as a real DD, it just happens to be able to use a lot of MNK DD gear but nothing NIN specific, so the numbers are approximate guesses. For calculations I'll use decimals even though the game rounds most of it down, 5 STR doesn't actually give 1.25 fSTR, it gives 2 25% of the time and 1 75% of the time, that's what the decimals represent here.

Base+fSTR 46/40, avg 43

5 STR means 1.25 fSTR and 2.5 attack

Berserk 60% of the time, 2.875 attack

2.875/350 = ratio increases 0.008, pdif too

44.25 * 1.608 = 71.154

43 * 1.6 = 68.8

3.42% more damage per hit with Flame

A sniper's is 5 acc, 2.5% effective acc

77.5% vs 75%

3.33% better acc with Sniper's

Those 2 increases result in basically the same DoT. If your ring happens to bump fSTR by 2 it'll end up winning by a moderate margin, if it bumps it by 1 then it'll lose by a smaller margin.

Of course, misses do not give TP so the accuracy difference also results in a slight TP gain slowdown. Lets say you need 21 hits to get 100 with your store tp amount. That means you'll need 28 swings with a Flame Ring on average, and 27.09 with a Sniper's. Basically, 1 more swing to 100% on average, on a double hit job with 15% DA assuming Brutal. So you'll need an extra round slightly less than half the time. A definite slowdown, but not exactly a huge one.

Replace Acc+5 with Acc+7 and the relative accuracy increase becomes 4.66, enough to create a more obvious gap. The whole point here is that even in a situation using attack food, the difference in overall performance is not always the humongous gap people like to picture. If base weapon damage is higher, effective acc is lower or fSTR is capped, obviously the balance will tip much further towards the accuracy ring. If accuracy is higher or capped, then it tips towards the STR ring instead.

Accuracy isn't going to cap on anything decently strong without sushi, but there are valid situations/reasons to want capped accuracy despite the fact that it is somewhat wasteful, so there are reasons to be in situations where the STR ring will result in strictly better overall performance.

So what's better in the ring slot varies with a number of variables, like anything else. Accuracy ring will win in the vast majority of situations where accuracy is uncapped, which does not actually mean every situation.

[aurik]

1.6 average pdif is absurdly high, which goes to show you can make up numbers that prove almost anything.

[Force-Attuned Krogoth]

Meat + bard, a pDIF of 1.6 is about right for mamools, especially considering he's talking about nin/war who should be berserked 60% of the time.

[Izzy]

(I realize that STR has a much larger role in most WSs).

Not true, acc still trumps STR on weaponskills, except where you're sneak-attacking (or trick-attacking with assassin trait) a one-hit WS.

Lets say for relic WS, (Specifically Randgirth(iknowwasteofcurrency)) wouldn't i wanna stack on the STR/modifiers instead of accuracy as it's a one hit WS?

[Nagamaki]

Absurdly high?

Against a lv82 mob with 350 def, 1.6 cRatio requires average 682 attack, with /WAR, attack food and 1 minuet than is indeed unrealistically high. I picked that cRatio somewhat randomly, it certainly could be achieved with a greater focus on NIN attack (merits, relic) and Dia II on the mob but I didn't use a reduced def number in there.

I'm overly used to MNK/WAR 450ish base attack and dia being applied to mobs, in which case a cRatio of 1.6 isn't very hard to achieve.

Instead, just take 1.6 as average pdif instead of cRatio, if you include crits in the average pdif 1.6 is achievable without requiring anything particularly crazy, for someone with good NIN DD gear/merits.

Either way, a lower pdif isn't going to change the base damage boost or change anything about accuracy, and will in fact mean that the tiny amount of attack on a STR ring will in fact represent a slightly larger relative pdif increase, so this doesn't exactly show anything thing about making up numbers to prove anything.

[aurik]

Absurdly high?

Against a lv82 mob with 350 def, 1.6 cRatio requires average 682 attack, with /WAR, attack food and 1 minuet than is indeed unrealistically high. I picked that cRatio somewhat randomly, it certainly could be achieved with a greater focus on NIN attack (merits, relic) and Dia II on the mob but I didn't use a reduced def number in there.

I'm overly used to MNK/WAR 450ish base attack and dia being applied to mobs, in which case a cRatio of 1.6 isn't very hard to achieve.

Instead, just take 1.6 as average pdif instead of cRatio, if you include crits in the average pdif 1.6 is achievable without requiring anything particularly crazy, for someone with good NIN DD gear/merits.

Either way, a lower pdif isn't going to change the base damage boost or change anything about accuracy, and will in fact mean that the tiny amount of attack on a STR ring will in fact represent a slightly larger relative pdif increase, so this doesn't exactly show anything thing about making up numbers to prove anything.

You're overly used to wearing +20 attack weapons, +20 attack body, +20 attack gloves with an A+ weapon; walking around with 450 base attack before food and songs is hardly "normal", especially for ninjas who wear less attack in all 3 of those major-boost slots.

My parses of extremely well geared & merited ninjas using meat with a single minuet put them typically somewhere in the 50s for average damage per swing, depending on conditions of course. I've got one parse easily accessible that showed over the course of a decent-length party a ninja averaging 57 damage per hit. With his high strength bonus (near capped at night due to relic hands, eg), I still calculated his pDIF average to be 1.25ish. Even if I'm lowballing it, that's a far cry from 1.6. BTW this was on Colibri with Dia2 on every single mob.

[razorcat]

Common sense leads me to believe that the faster you TP (less misses, more attack rounds from haste) leads to more you WS. More WS means that more damage is done to a monster which usually means it dies faster. Faster monster deaths generally means faster xp rate. Whats so hard about that logic?

[aurik]

Common sense leads me to believe that the faster you TP (less misses, more attack rounds from haste) leads to more you WS. More WS means that more damage is done to a monster which usually means it dies faster. Faster monster deaths generally means faster xp rate. Whats so hard about that logic?

A majority of damage dealt to the monster is done during TP phase rather than WS.

[Septimus]

Common sense leads me to believe that the faster you TP (less misses, more attack rounds from haste) leads to more you WS. More WS means that more damage is done to a monster which usually means it dies faster. Faster monster deaths generally means faster xp rate. Whats so hard about that logic?

Since everyone is nitpicking in this thread, it is possible to kill mobs too fast. If you are killing a pool of mobs faster than they can repop, you will lose chain and your XP/hour will drop.

[Amele]

Common sense leads me to believe that the faster you TP (less misses, more attack rounds from haste) leads to more you WS. More WS means that more damage is done to a monster which usually means it dies faster. Faster monster deaths generally means faster xp rate. Whats so hard about that logic?

Since everyone is nitpicking in this thread, it is possible to kill mobs too fast. If you are killing a pool of mobs faster than they can repop, you will lose chain and your XP/hour will drop.

god I hate that when I'm on a backline job. it's like asking a melee to saw off their own leg to kill things slower or something, I swear.

[Nagamaki]

For non crit pdif to average 1.25, cRatio would be around 1.25, maybe 1.3 since the mode drags average down slightly from what I've seen, although pdif's distribution seems to compensate for it a bit so the average ends up pretty close to cRatio despite the 1.0 mode.

Lets just go with 1.3.

Lv81 Colibri, Def 322.
Dia IId, 289.
dLv of 6 so cRatio 1.3 requires ratio 1.6
1.6 * 289 = 462

Lv82 Colibri, Def 327. Dia IId, 294.
dLv of 7 so cRatio 1.3 requires 1.65 ratio
1.65 * 294 = 485

Assuming even level distribution, that means 473.5 attack average.

Colibris so the meat would be something cheap, lets go with hobo food NQ miths. 65 attack 5 str so 67.5, attack pre-food 406.

Factor Berserk 60% of the time, 406 / 1.15 = 353

Substract Minuet IV, 287 base attack. Clearly, this isn't right. I haven't done massively extensive tests on pdif behavior, but previous tests on various mobs did show cRatio behaved pretty much how the simple equation says it should, listed attack divided by /c defense and corrected by 0.05 per level of difference. It might be slightly off, but not this off.

Dia II obviously doesn't always get on the mob before anyone hits it, sometimes minuet will be down if single BRD or the minuet BRD is pulling, and I guess Berserk might not be used as soon as it wears for some reason, but even all of those factors shouldn't lower average attack over time to nearly that level.

[aurik]

Time for some math without numbers.

O = D * P * A + W * F
O is overall expected damage per hit
P is pdif average
A is hit percentage
D is base damage
W is weaponskill damage
F is weaponskill frequency

The assumption is that W is constant, due to macro-switching (ie, you can always TP in 1 gearset and macro to different WS gear)

If you add strength, you get a resulting dP and dD, the amount of increase
dO(str) = (dD+D)(dP + P)A + WF
dO(str) = DPA + dDPA + DdPA + dDdPA + WF

If you add accuracy, you get a resulting dA and dF, the amount of increase
dO(acc) = DP(dA+A) + WF + WdF

Let's find the balance point

dO(acc) = dO(str)
WF+ WdF + DPA + DPdA = DPA + dDPA + DdPA + dDdPA + WF
WdF + DPdA = A(dDP + DdP + dPdD)
WdF / DPA + dA/A = (dDP + DdP + dPdD)/DP
WdF / DPA + dA/A = dD/D + dP/P + dPdD/DP
dPdD/DP will always be almost insignificant, so we can simplify to
WdF / DPA + dA/A = dD/D + dP/P

Let's plug in some "imaginary" numbers.

Pulling from a rdm rdm brd nin mnk thf parse on Colibri (dia2 on every monster) the ninja (s/u) had:
D=45; dD = 1.25
P=1.27; dP = .008
A=.708; dA = .025 (nq sniper)
W=544; dF = ?

544dF / 40.5 + .025/.708 = 0.0278 + .00630
13.4 * dF + .0353 = .0341
dF = -.0012 / 13.4

The conclusion here is that a sniper is equivalent to a flame ring if the sniper ring strictly decreases the frequency of using a weaponskill by 9 in 1000. Obviously, sniper rings have a positive effect, so sniper ring outperforms.

Using Nagamaki's imaginary numbers:
D=43; dD = 1.25
P=1.6; dP = .008
A=.75; dA = .025 (nq sniper)
W=544; dF = ? (I'm filling in a W here since he didn't guess one for us)

544dF / 40.6 + .025/.75 = 0.0291 + .005
13.4*dF + .0333 = .0341
dF = .0008/13.4 = .0000597

The conclusion here is that a sniper is equivalent to a flame ring if the sniper ring increases the frequency of using a weaponskill by 6 in 1000. This is clearly the case (it should increase the rate by at least 20 in 1000, I'd wager)

Note that these conditions (low base-damage, low WS damage) are ideal for boosting strength. In a situation with axes, hand-to-hand, or 2-handed weapons, adding accuracy is a largely superior choice.

[aurik]

For non crit pdif to average 1.25, cRatio would be around 1.25, maybe 1.3 since the mode drags average down slightly from what I've seen, although pdif's distribution seems to compensate for it a bit so the average ends up pretty close to cRatio despite the 1.0 mode.

Lets just go with 1.3.

Lv81 Colibri, Def 322.
Dia IId, 289.
dLv of 6 so cRatio 1.3 requires ratio 1.6
1.6 * 289 = 462

Lv82 Colibri, Def 327. Dia IId, 294.
dLv of 7 so cRatio 1.3 requires 1.65 ratio
1.65 * 294 = 485

Assuming even level distribution, that means 473.5 attack average.

Colibris so the meat would be something cheap, lets go with hobo food NQ miths. 65 attack 5 str so 67.5, attack pre-food 406.

Factor Berserk 60% of the time, 406 / 1.15 = 353

Substract Minuet IV, 287 base attack. Clearly, this isn't right. I haven't done massively extensive tests on pdif behavior, but previous tests on various mobs did show cRatio behaved pretty much how the simple equation says it should, listed attack divided by /c defense and corrected by 0.05 per level of difference. It might be slightly off, but not this off.

Dia II obviously doesn't always get on the mob before anyone hits it, sometimes minuet will be down if single BRD or the minuet BRD is pulling, and I guess Berserk might not be used as soon as it wears for some reason, but even all of those factors shouldn't lower average attack over time to nearly that level.

Here's the parse I'm referring to:

http://img440.imageshack.us/img440/9...07parseqh6.png
thf/war (mandau) mnk/war (spharai) nin/war (s/u), rdm rdm brd/nin, sis kebabis, minuet, march. Approx 80% colibri 20% bugard, always debuffed and str-down erased promptly.

Even as mnk/war, with a guesstimated base damage of 65, I'm averaging 1.5ish -- which is deliberately overestimating because kick attacks (melee/dune/5 merits) and triple-procs factor into the final average (96) while having a higher base damage.

A ninja using anything other than kikoku, and using dusk/koga and haub will have far less attack than I did, I don't see how they can average 1.6.

[evilpaul]

As a decently gear NIN with 8/8 merits in Katana I usually have ~385 Attack in TPing gear, ~410 on WS.

Your numbers are mostly a mystery to me, but I'm not sure what I could wear to have ~450 Attack before food/Berserk short of a Kikoku, TPing in O.kote instead of Dusk, etc. Particularly while maintaining a decent accuracy % on Greater Colibri.

[Nagamaki]

450 is MNK/WAR attack range, that's not really was I was thinking for NIN base, and its not really required with debuffs and all applied.

A parse that includes the Bugards there will certainly have its numbers thrown off quite radically, since Bugards have a greater defense rating, greater VIT and an additionnal 51/256th defense multiplier to their natural def, on top of a powerful STR down and defense bonus buff. The STR down and defense bonus can be removed quickly to have a fairly small impact, but the natural high defense of bugards is always there. Dia II and an Acid Bolt effect combined would roughly counteract the racial trait and bolts did work pretty well on those guys, no mention if THF was using them or a bomblet though.

With Senj/Unji and constantly capped fSTR that NIN would have have an avg cRatio around 1.25, fSTR doesn't cap that easily even with Koga at night though, and certainly not on the bugards, without stacking a lot of STR that probably wasn't there, so fSTR was most likely not capped which would pump that cRatio up a bit. Combine that with the bugards and getting the real average ratio on the GCs is really difficult.

Taking the example that someone posted with 385 base attack, a NIN with 1 minuet, 60% serk and Sis Kebabi would have 591 average attack, average cRatios of 1.73 and 1.66 on lv81 and 82 diad GCs. The max non-crit of 100 suggests the NIN's base attack gear might have been slightly lower along with fSTR probably never reaching the cap for Senj (which would have meant 50 base). You need 111 STR to get fSTR to 12 (Senj cap) on GCs, that requires a fair amount of STR gear even as Elv with STR merits (80 base), certainly more than a NIN is likely to have for meleeing even at night. Actually, the highest crit for the NIN is 147, so fSTR probably capped out at 11, for 49 base damage, and presumably fSTR 8 during the day. The lows for everybody in that parse are really really low, that's another sign of the bugard throwing the numbers off.

[aurik]

450 is MNK/WAR attack range, that's not really was I was thinking for NIN base, and its not really required with debuffs and all applied.

A parse that includes the Bugards there will certainly have its numbers thrown off quite radically, since Bugards have a greater defense rating, greater VIT and an additionnal 51/256th defense multiplier to their natural def, on top of a powerful STR down and defense bonus buff. The STR down and defense bonus can be removed quickly to have a fairly small impact, but the natural high defense of bugards is always there. Dia II and an Acid Bolt effect combined would roughly counteract the racial trait and bolts did work pretty well on those guys, no mention if THF was using them or a bomblet though.

With Senj/Unji and constantly capped fSTR that NIN would have have an avg cRatio around 1.25, fSTR doesn't cap that easily even with Koga at night though, and certainly not on the bugards, without stacking a lot of STR that probably wasn't there, so fSTR was most likely not capped which would pump that cRatio up a bit. Combine that with the bugards and getting the real average ratio on the GCs is really difficult.

Taking the example that someone posted with 385 base attack, a NIN with 1 minuet, 60% serk and Sis Kebabi would have 591 average attack, average cRatios of 1.73 and 1.66 on lv81 and 82 diad GCs. The max non-crit of 100 suggests the NIN's base attack gear might have been slightly lower along with fSTR probably never reaching the cap for Senj (which would have meant 50 base). You need 111 STR to get fSTR to 12 (Senj cap) on GCs, that requires a fair amount of STR gear even as Elv with STR merits (80 base), certainly more than a NIN is likely to have for meleeing even at night. Actually, the highest crit for the NIN is 147, so fSTR probably capped out at 11, for 49 base damage, and presumably fSTR 8 during the day. The lows for everybody in that parse are really really low, that's another sign of the bugard throwing the numbers off.

As I noted before, the share of bugards was low, the str-down was removed promptly as was the bugard defense buff. This isn't your typical pickup pt, as you might imagine from the particular setup and weapons involved.

You should always ignore lows on parses because there's always going to be the occasional stray low-pdif hits due to factors. For example, if a colibri eats the food as it dies, and a bugard is autotargetted, one might swing before reapplying food and dia2 is cast. Add to that an unlucky roll with pdif and using unji, and getting the stray 20dmg hit is not unreasonable, and it only has to happen once to show up.

However, you don't get a concept of how frequent the low values were reached, so the "low" data is meaningless. Only the high data and mean data give us information that is usable, and in the case of monks, thfs, and relic users, the "highs" are also corrupted by various factors.

As you can see from the parse, the high crit for the nin is 147; assuming pdif of 3, this is a fstr of 11 at night, 8 during the day, leading to katana base of 49/43 (46 average) during the night, 46/40 (43 average) during the day. The true average over the course of the party then should be between 43 and 46 on the easiest mob.

However, even choosing 42 or even 41 to be the basedamage average over the course of the party (which is unlikely considering the frequency of bugards) the pdif average is still not much higher than 1.35. Which leads to the question, how can you justify a 1.6 average on mobs that don't have defense as low as colibri?

[Necrosis]

On the topic of acc, my meat TP set works well for me; I usually don't have acc problems anywhere (on THF). However, I'm wondering what THFs use for a WS set @ camps like Mamool Ja Staging point. That is, what mix of acc/str/dex do you use on non-puk mobs?

[Nagamaki]

I already stated that the 1.6 figure was too high to be reasonable in the exact situation I made up, so hammering on that is pretty pointless. It was a somewhat random number I made up without calculating the gear implied with the amount of def. It has no real impact on what the comparison even was in the first place.

But since you insist, lets take a very high level camp, Ulli's room.

Lv81
349, Dia IId to 314
1.9, needs 596

Lv82
354, Dia IId to 318
1.95, needs 620

Lv83
359, Dia IId to 323
2.00, needs 646

Lv84
365, Dia IId to 328
2.05, needs 672

Avg with even level distribution to hold 1.6 cRatio avg: 633.5

With 385 base attack, min IV, serk, h.pie, avg attack of 608, which falls a bit short. You could achieve it if a separate def down is applied regularly, or by using dragon food, but those aren't really common. These are the cRatios with 608 attack :

1.636
1.561
1.482
1.403

Avg cRatio 1.520 at a camp with a greater max level and average def than any camp people actually merit without doing anything all that special. 1.6 cRatio, going by the accepted math involved in calculating it, isn't particularly hard to achieve if you put the same stats on a NIN fighting mobs with a slightly lower average level.

If the Bugards didn't drive average pdif down in that parse, exactly what factor justifies the average being so much lower than the result obtained through applying the formulas using known stats? Is the formula simply grossly inaccurate? Sadly the Bugards are gone and I never recorded their def, but its still clear from their known trait that it was significantly higher than that of the Colibris.

[aurik]

If the Bugards didn't drive average pdif down in that parse, exactly what factor justifies the average being so much lower than the result obtained through applying the formulas using known stats? Is the formula simply grossly inaccurate? Sadly the Bugards are gone and I never recorded their def, but its still clear from their known trait that it was significantly higher than that of the Colibris.

I think that the formula isn't entirely accurate.

Doesn't the formula make the assumption that the distribution of pDIF selection among the min and max bound is uniform? Because I know that is not the case; the chance of hitting for exactly base damage is inordinately high at certain regions on the atk/def ratio line.

For example, this data is from a KRT party from early 2006. I was using Destroyers back then.
This is the number of occurences of normal melee hits against Spartoi Warrior from me for the duration of the party (um, about 30k xp?). Notice the huge frequency spike at my current base damage, while the rest of the normal spectrum is roughly flat.

Aurik

NORMAL

56	2
60	1
61	8
63	83          <<<<<<<<<<<<<<<<<<<<<<<<<
64	9
65	14
66	20
67	17
68	11
69	9
70	20
72	7
73	12
74	17
75	14
76	29
77	7
78	19
79	11
81	15
82	25
83	11
84	23
85	13
86	23
87	20
88	33
90	14
91	27
92	27
93	19
94	23
95	19
96	26
97	29
99	26
100	16
101	16
102	23
103	19
104	16
105	28
106	14
108	26
109	16
110	26
111	22
112	18
113	13
114	17
115	7
117	6
118	15
119	12
120	19
121	14
122	15
123	12
124	9
126	16
127	9
128	12
129	9
130	15
131	16
132	10
133	6
135	5
136	4
137	7
138	2
139	4
140	2
141	1
142	2
144	1
145	1
148	2
149	1
150	1
154	1
156	1
159	4
160	1
165	2
167	1