Best in Slot for SMN after 2.1?

[T0rin]

That's the same logic for the blm spell speed set, but as a smn it is very rare for me to have to break a cast, we have short casts with shadow flare being the only one over 2.5. Maybe while learning fights having a spell speed set would be useful.

There are certain points in the Titan Ex fight where I know Miasma will finish only if I really risk a landslide. There are some times I break a ruin while dodging, where if I just had a smidgeon of spell speed, I'd have finished it. Some of the time you really just can't afford to risk the potential for lag on mechanics, but the overall benefit of SS is so minor and subjective that it is hard to quantify, especially when it comes to BiS.

[Madoka]

Generally if I had to break a cast to dodge I could use Ruin 2 while on the move. But I've also gotten pretty used to exactly how many casts I can get off before I need to move during fights like ex primals and coil. Like I can use RS, drop all 3 dots and hit contagion/fester as I dodge Titan's first landslide in EX, pop a shadowflare and be in position to stack/dodge the first set of weights, start ruining and drop virus just before the first MB (lasts for the tumults after if you get it right) and keep going from there.

Usually my speed at dodging/positioning determines if I get two casts instead of one when dodging and my SS hasn't changed much since I started the fights. Actually my SS has probably dropped if anything going into i90 and I'm experiencing the same amount of successful casts as ever.

[jars]

the ultima ring seems worse than an overmelded astral ring. that's disappointing

[SolarMisae]

Edit: Definitely more than 1 item is changing. :P Went back to experiment more with other BiS potential sets, I had overlooked some...

Nooo. So many changes. D: I just barely completed the old BiS by getting the belt...(minus t5 book...myeh...) and now this. ;w;

Such sad. At least I can wear my sexy Summoner pants (or lack-thereof) again. :x

But now I need to hope for allagan gloves if I go with set 2, so I don't have to worry about the ring...

[Zheri]

T0rin, I'm also having some trouble with your calculations and was hoping you or someone else who understands the numbers better than me could offer insight. This week I'm in a position to buy either the +1 relic or pants* and tried running the numbers through a spreadsheet to see which one is actually a bigger upgrade. Accounting for base stats and food, the +1 gives me a crit-modified base damage value of 105.467958 and a weighted int of 1072.667559. Buying the pants gives me a base damage of 105.244842 and weighted int of 1093.005167. So I have a higher damage with the +1, yet higher weighted Int with the pants. I know that SS is not a factor since I actually lose 17 by going with the pants. Just to double check myself, I also ran the same numbers you used through it (71 WD, 499 Int, 262 Det, 565 Crit, and 378 SS) and got the same results as you. I'm curious as to how this is possible in the first place, and which purchase would actually benefit me more.

For reference, here's the numbers I used
________WD_____Int_____Det_____Crit_____SS
Control___71_____499____262_____565_____378
Relic +1__69_____461____319_____447_____413
Pants____66_____475____315_____470_____396

*Normally, the +1 would be a no-brainer, but due to accuracy issues I'm forced to use a few i70 pieces in place of i80/90 that I have from gearing BLM.

[T0rin]

snip

For starters, ignore the int weighting for now, there is a flaw in the logic that causes a bias towards gear with certain stats, and isn't a true representation of the dps potential of the gear. The crit modified damage value is probably the easiest and quickest measure for figuring out a best single upgrade.

In your case... the relic +1 gives a value of 105.5370734 and the pants give a value of 105.2448423. Even though you have to use i70 in place of 80/90 to maintain accuracy, the relic is still a better upgrade. If you were in a more 'normal' situation, i.e. someone who had full i70, going from relic to relic+1 or darklight pants to myth pants, the upgrade choice would be that much easier to make.

So, in terms of DPS, the Relic +1 is a better upgrade. But, a better question is this: Can you get the primal weapon? (Omnitome) If so, that would be a better choice. It's good to help get through the SMN accuracy issues transitioning out of i70 gear, and would allow you to wear your i80/90 BLM pieces, and have myth available. If it was around when I was going from i70 to i90, I'd have gladly used that thing leading up to getting the allagan weapon. But, if that isn't a possibility, the Relic is a better upgrade in the immediate terms.

For kicks, I ran your data through my sim to see which would work best, here's what I got:

Simulation complete - stats = WD: 69, INT: 461, DTR: 319, CRT: 447, SS: 413, GCD: 2.43, PGCD: 2.93, total = 60211 damage
Simulation complete - stats = WD: 66, INT: 475, DTR: 315, CRT: 470, SS: 396, GCD: 2.45, PGCD: 2.95, total = 58750 damage

[Zheri]

...snip...

Thank you. I appreciate the help. Hopefully I can get the Omnitome this week making the whole debate moot, but I'm not holding my breath.

[Zheri]

For starters, ignore the int weighting for now, there is a flaw in the logic that causes a bias towards gear with certain stats, and isn't a true representation of the dps potential of the gear. The crit modified damage value is probably the easiest and quickest measure for figuring out a best single upgrade.

Sorry for the double post, but I actually see what was wrong in the calculations. The entire concept of Weighted Int, as presented, is flawed. Because the absolute value of Int scales with Crit, WD, and to some degree Spell Speed, we would have to choose a gear set and let its delta(Int) be the standard by which we determine stat weights, rather than weighing stats relative to the gear set where 1 Int is always equal to 1 Weighted Int.

Going back to the spreadsheet I made yesterday, I set delta(Int) = 0.1579255, or the value returned using the stats listed under Control, and recalculated. The new totals including base stats and food came out to be slightly more than 9 points in favor of the +1. This agrees with both the base damage and the sim, which makes me happy for now. The biggest difference is that now 1 Int =/= 1 Weighted Int.

Giving your "Old BiS" and "New BiS" the same treatment actually produced interesting results, putting the old set on top, which is in line with the crit-adjusted damage values.
"Old" Weights: WD = 6.0490022 . Int = 1.00315465 . Det = 0.2820812 . Crit = 0.1788873 . SS = 0.08943755 . Total = 714.5100490
"New" Weights: WD = 6.0473119 . Int = 0.09807566 . Det = 0.2829711 . Crit = 0.1840486 . SS = 0.08943755 . Total = 709.7806269

[T0rin]

snip

What stats did you use to calculate the delta int to use in re-adjusting the weights?

You know, the more I think about it, the less value I think there is in using stat weights. Let's stop to think for a second... what does a stat weight really mean?

It is a _relative_ value of a single stat when increased above a baseline value.

What we know about the damage formula is pretty simple: the weight of INT, WD and CRT change with the baseline values. This basically means, that when you take a baseline, and create weights, those weights are only valid if all you do is change a single piece of gear that adjusts your baseline stats by 1 in total. After that, you need to calculate a new set of weights.

So, when you create a baseline that includes all stats in their 'final' (no further upgrades will be applied) form, all you're doing is finding out the current worth of each stat, were you to actually increase any of them by 1, which if this is your actual 'final' set of gear, you won't be. If you take those weights and multiple your current gear contributions by them, all you're doing is finding the value of that using current weights, if you were to apply it all over again, which is something you would never actually do.

Basically, since any weight is really just a value relative to the current baseline if you were to increase a stat by 1, there is no 'magic number' that you can apply to an entire set of gear to find out their worth.

Even if you were to take a baseline from your completely 'naked' state, the weights would not be reliable. Since the relative values of WD, INT and CRT are assuredly going to change as you equip gear, you can't use your pre-gear or post-gear weights to attribute 'value' to the gear set as a total. And since the 'value' of a single point of most stats is relative to your other stats, even a single specific item may be worth more or less to one person, than another. If two people have two different sets of gear, but lack an earring, for instance... one set of gear is DTR heavy and the other set of gear is CRT heavy, the benefit each person will receive from the same piece of gear will be different. If that earring were to be heavy on CRT, the actual contribution to damage will be higher for the DTR heavy gear set, but if the earring were to be heavy on DTR, the damage contribution would be higher for the CRT heavy gear set. Even in this simple example, we can see how there is no such thing as a stat weight that can be used to compare the value of gear, as the value of any piece will be different based on current stats.

I think that due to the variable nature of stat values given the damage formula, we can never really come up with an INT weight or 'stat score' that could be attributed to 1 piece of gear, much less a whole set of gear. At this point, I think the only real way to compare 2 sets of gear is their calculated crit modified damage values, or more accurately, a simmed rotation that uses those values.

And even a sim is only worth so much, because it basically represents doing DPS in a bubble, with no interruptions. And due to the nature of the value of spell speed, those sims might actually need to take into account a certain level of spell interruption (like a Ex Titan fight or whatever), mob availability, (Titan jumps) etc. Basically, no data we come up with will be perfect, but if we can sim out a semi-realistic situation, we can get a good picture of how well a set of gear stacks up against another.

[Thorauku]

Combo breaker for awhile.

Does any parser actually parse the damage of Summoner + Pet? I mean the merged of it?