I take a day off and another page of posts rolls by, le'see...

Originally Posted by
Shinryuu
What server are you on, Araelus? And if you're not interested in parsing I might go do it, the Grand Marid thing is interesting if they really only spawn as level 80. They'd become the new Ronfaure bunnies, so to say.
I'm on .... that's right, Quetzalcoatl. As for parsing, lately I've been tied up with a lot of real-world business, so go for it.
If you get different xp results from two kills, though, no dice.

Originally Posted by
pchan
Didn't read your charts, the things I punched are a mixed of imps and tough qiqrn which, last I checked don't have awesome defense, it's more the opposite... Look like someone has cuchulain's mantles for sale or someone's charts sucks.
Looks like a false dichotomy! I have no Cuch Mantle, if I had one I would not be selling it, and I am confident in the charts, even the 100% damage from TP charts. I don't know if you care, but the crunchy numbers show that as a monster's defense is lower, cēterīs paribus, an efficient agent will add accuracy to cap before aiming for even low values of attack.

Originally Posted by
Shinryuu
Araelus, you're treating scum like pchan way too nice.
Either Donald Davidson's principle of conversational charity, or Matthew 5:43–48, same result.

Originally Posted by
Wolfknight
sorry, I meant if acc is 5 times better than attack why the hell would you even bother using meat is my point, when sushi or pizza give acc.

Originally Posted by
Mayareira
Baseless magical(!?) numbers like "5 times better", "2 att > 1 acc" won't get people anywhere. Rule of thumb is that sacrificing x amount of accuracy(x% of your hitrate) for whatever amount of attack is generally a bad idea. I could get into the math stuff, but shits got to be common sense already and there is simply no point wasting time on it.
Mayaerira has is exactly right, Wolfknight. The point of those charts is to show that the equivalent amounts of attack and accuracy during TP, vs. a single monster, change depending on your current hit% and Attack.
As a person adds +ATT or +ACC, each additional point has decreasing marginal returns. The intuitive move is to add as much accuracy and haste using gear as you can, then eat Attack food, except when +ATT adds more damage than +ACC.
What surprised me was how the math shows it is very difficult to increase your total damage more by adding +1 attack than by adding +1 accuracy, especially on merit mobs.
This is what you're going to have to deal with, Wolfknight: it is "magical thinking" to have +X attack equal to +Y accuracy as some kind of constant rule. And the more of your total damage that comes from WS, the more the balance swings against +ATT during TP.

Originally Posted by
Shinryuu
"So the sun doesn't orbit around the earth, that seems really extreme to me, I could see the sun not orbiting around the moon but the sun not orbiting around the earth seems nuts even with telescopical proof"
Yeah, exactly! This is a great example: we need to be careful what ideas we make into dogmatic truths. Now Shaddix can take this and turn it around on me in terms of "+1 acc = ??? hitrate", so I will appreciate any and all testing on that front.

Originally Posted by
Max™
The key thing to consider is, how much does 1 miss hurt, compared to a swing landed with more attack.
That is what Araelus' chart was trying to display, the points where you lose more by missing a hit, than you lose hitting for slightly less.
Apparently the chart format is still not easy for everyone to use - please let me know what I can do to make it more use-able.
For me it is intuitive: find current Accuracy and Attack, then add to the stat which is farther to the right, moving left along that stat line as the value increases, and at every point it shows you how much dmg+% the att/acc gives.

Originally Posted by
Bakthi
Araelus: I really liked the charts way back on page 3 comparing the benefits of adding acc/attack. Any chance of throwing haste in there, or just doing haste versus acc? I typically TP in a haste setup in merits on COR, and my acc is reasonable there, but tends to be far less reasonable in Einherjar, and I'm curious about the benefits of haste versus accuracy given various hit rates/haste totals.
Bakthi, I am glad you could use the charts.
I can add haste in, but I need to know how to compute the actual increase in damage before I do so.
My problem is that Haste+ has increasing marginal returns. The "easy" answer is to say, cap hitrate before adding Haste, but depending on your total Haste (not just gear but also spells and songs) there are some points where it is more valuable to be "missing faster" than "hitting slower" if that makes sense.
The Haste+ gained besides equipment is my roadblock. I will try to make more than one Haste% curve, showing situations that might arise. However, since I am currently working on my tax returns (how do I owe taxes when I lost money overall this year?!) my brain is acquiring the consistency of week-old porridge.
And so I ask the math gurus:
Code:
delay delay delay delay delay delay
haste delay change | haste delay change | haste delay change | haste delay change | haste delay change | haste delay change
0 450.0 0.00% | 0 450.0 | 0 450.0 | 0 450.0 | 0 450.0 | 0 450.0
1 445.5 -1.00% | spell 15 382.5 | spell 15 382.5 | spell 15 382.5 | spell 15 382.5 | spell 15 382.5
2 441.0 -1.01% | 16 378.0 -1.18% | hasso 25 337.5 | m2 26 333.0 | m2 26 333.0 | m2 26 333.0
3 436.5 -1.02% | 17 373.5 -1.19% | 26 333.0 -1.33% | 27 328.5 -1.35% | hasso 36 288.0 | hasso 36 288.0
4 432.0 -1.03% | 18 369.0 -1.20% | 27 328.5 -1.35% | 28 324.0 -1.37% | 37 283.5 -1.56% | march 45 247.5
5 427.5 -1.04% | 19 364.5 -1.22% | 28 324.0 -1.37% | 29 319.5 -1.39% | 38 279.0 -1.59% | 46 243.0 -1.82%
6 423.0 -1.05% | 20 360.0 -1.23% | 29 319.5 -1.39% | 30 315.0 -1.41% | 39 274.5 -1.61% | 47 238.5 -1.85%
7 418.5 -1.06% | 21 355.5 -1.25% | 30 315.0 -1.41% | 31 310.5 -1.43% | 40 270.0 -1.64% | 48 234.0 -1.89%
8 414.0 -1.08% | 22 351.0 -1.27% | 31 310.5 -1.43% | 32 306.0 -1.45% | 41 265.5 -1.67% | 49 229.5 -1.92%
9 409.5 -1.09% | 23 346.5 -1.28% | 32 306.0 -1.45% | 33 301.5 -1.47% | 42 261.0 -1.69% | 50 225.0 -1.96%
10 405.0 -1.10% | 24 342.0 -1.30% | 33 301.5 -1.47% | 34 297.0 -1.49% | 43 256.5 -1.72% | 51 220.5 -2.00%
11 400.5 -1.11% | 25 337.5 -1.32% | 34 297.0 -1.49% | 35 292.5 -1.52% | 44 252.0 -1.75% | 52 216.0 -2.04%
12 396.0 -1.12% | 26 333.0 -1.33% | 35 292.5 -1.52% | 36 288.0 -1.54% | 45 247.5 -1.79% | 53 211.5 -2.08%
13 391.5 -1.14% | 27 328.5 -1.35% | 36 288.0 -1.54% | 37 283.5 -1.56% | 46 243.0 -1.82% | 54 207.0 -2.13%
14 387.0 -1.15% | 28 324.0 -1.37% | 37 283.5 -1.56% | 38 279.0 -1.59% | 47 238.5 -1.85% | 55 202.5 -2.17%
15 382.5 -1.16% | 29 319.5 -1.39% | 38 279.0 -1.59% | 39 274.5 -1.61% | 48 234.0 -1.89% | 56 198.0 -2.22%
16 378.0 -1.18% | 30 315.0 -1.41% | 39 274.5 -1.61% | 40 270.0 -1.64% | 49 229.5 -1.92% | 57 193.5 -2.27%
17 373.5 -1.19% | 31 310.5 -1.43% | 40 270.0 -1.64% | 41 265.5 -1.67% | 50 225.0 -1.96% | 58 189.0 -2.33%
18 369.0 -1.20% | 32 306.0 -1.45% | 41 265.5 -1.67% | 42 261.0 -1.69% | 51 220.5 -2.00% | 59 184.5 -2.38%
19 364.5 -1.22% | 33 301.5 -1.47% | 42 261.0 -1.69% | 43 256.5 -1.72% | 52 216.0 -2.04% | 60 180.0 -2.44%
20 360.0 -1.23% | 34 297.0 -1.49% | 43 256.5 -1.72% | 44 252.0 -1.75% | 53 211.5 -2.08% | 61 175.5 -2.50%
21 355.5 -1.25% | 35 292.5 -1.52% | 44 252.0 -1.75% | 45 247.5 -1.79% | 54 207.0 -2.13% | 62 171.0 -2.56%
22 351.0 -1.27% | 36 288.0 -1.54% | 45 247.5 -1.79% | 46 243.0 -1.82% | 55 202.5 -2.17% | 63 166.5 -2.63%
23 346.5 -1.28% | 37 283.5 -1.56% | 46 243.0 -1.82% | 47 238.5 -1.85% | 56 198.0 -2.22% | 64 162.0 -2.70%
24 342.0 -1.30% | 38 279.0 -1.59% | 47 238.5 -1.85% | 48 234.0 -1.89% | 57 193.5 -2.27% | 65 157.5 -2.78%
gear 25 337.5 -1.32% | 39 274.5 -1.61% | 48 234.0 -1.89% | 49 229.5 -1.92% | 58 189.0 -2.33% | 66 153.0 -2.86%
spell 40 270.0 | gear 40 270.0 -1.64% | 49 229.5 -1.92% | 50 225.0 -1.96% | 59 184.5 -2.38% | 67 148.5 -2.94%
m2 51 220.5 | m2 51 220.5 | gear 50 225.0 -1.96% | gear 51 220.5 -2.00% | 60 180.0 -2.44% | 68 144.0 -3.03%
hasso 61 175.5 | hasso 61 175.5 | m2 61 175.5 | hasso 61 175.5 | gear 61 175.5 -2.50% | 69 139.5 -3.12%
march 70 135.0 | march 70 135.0 | march 70 135.0 | march 70 135.0 | march 70 135.0 | gear 70 135.0 -3.23%
Is that right? (Gear +25%*, Spell +15%, March II +11%, March +9%, Hasso +10%) *must have +26% in gear to get 25% haste, hurr durr.
The important bit is that the more Haste you have before gear is taken into account, the more that each +1% haste from gear does.
Starting from 0% Haste, the first +1 gives you -1.00% delay, and the final point of gear Haste will give you -1.32% delay.
Starting from 45% Haste, the first +1 gives you -1.82% delay, and the final point of gear Haste will give you -3.23% delay.
Shown graphically:
And once again, I do not know why imageshack refuses to host these chart images. It keeps getting auto-removed. 
So, Bakthi, if it helps - you can see on this chart #8 how much each point of haste from gear will reduce your attack speed. You can compare that to the effect of Accuracy, and it's not too bad, although they are not on the same chart... hmm.
Haste+1% at zero Haste: -1% delay
Haste+1% at 69% Haste: -3.23% delay
Acc+1 at 20% hitrate: damage +2.5%
Acc+1 at 94.99% hitrate: damage +0.01%
Att+1 when pDIF(avg) goes from negative to positive: damage +5%
Att+1 later on: no effect.
In other words:
During TP phase!
Haste will always do something until capped, and it does more and more on the margin until capped.
Att+1 does almost nothing until pDIF(avg) gets over zero, because while pDIF(max) is never negative, pDIF(min) can be - so at the point where Att+1 brings you up above pDIF(avg)=0 it is particularly amazing, but then Attack quickly loses efficiency.
Acc+1 when at the lowest possible hitrate will increase your damage by 2.5%, and when you are at 94% hitrate it still increases your damage by 0.53%.
TL;DR
Haste +1% ≤ -3.23% delay, starts at -1% delay, but increasing returns
Attack+1 ≤ DMG+5%, returns decrease very fast
Accuracy+1 ≤ DMG+2.5%, returns decrease very slowly