RDM thoughts and tuning needed

[Leidiriv]

If you're going to look at it that way (and I recommend we do), you also need to account for the mana that would have been generated by the spells you're not casting during the combo. This puts the effective cost of a full melee combo at about 163 mana, as the melee combo expends 160 and generates 35 in the time that you could have generated 38 mana by casting spells.

The problem with looking at the generation is that if you look at the generation in the melee combo, it's all hypothetical mana because you will overcap in that case. Mana wasted on overcapping has a potency of 0. Additionally, the reason we consider mana that could have been generated in the time spent on a weaponskill is because that mana could have been built toward a melee combo, and you can't build very far at all toward another melee combo in place of doing the melee combo.

[Rongway]

The problem with looking at the generation is that if you look at the generation in the melee combo, it's all hypothetical mana because you will overcap in that case. Mana wasted on overcapping has a potency of 0. Additionally, the reason we consider mana that could have been generated in the time spent on a weaponskill is because that mana could have been built toward a melee combo, and you can't build very far at all toward another melee combo in place of doing the melee combo.

There exists a mana range (20-180) where you could either execute a weaponskill or cast spells. That those weaponskills are the wrong choice during much of this range is irrelevant; they are usable. There's also a range (160-180ish) where you could execute a combo or continue to cast spells. Because of this, all that "hypothetical" mana is real mana.

[Leidiriv]

There exists a mana range (20-180) where you could either execute a weaponskill or cast spells. That those weaponskills are the wrong choice during much of this range is irrelevant; they are usable. There's also a range (160-180ish) where you could execute a combo or continue to cast spells. Because of this, all that "hypothetical" mana is real mana.

Okay, say we're at 200 mana (or, since we're not barbarians and want to guarantee the Stone/Fire proc off Holy/Flare, 199). We can't possibly generate any more mana. Does the melee combo still have the extra "cost" of the 28-35 mana you could have generated in that time?

[Rongway]

Okay, say we're at 200 mana (or, since we're not barbarians and want to guarantee the Stone/Fire proc off Holy/Flare, 199). We can't possibly generate any more mana. Does the melee combo still have the extra "cost" of the 28-35 mana you could have generated in that time?

That scenario isn't worth considering until we've already assigned value to a point of mana. What we are interested in are question like

• Should I spend my mana on E.Reprise or should I spend it on E.Riposte...Scorch?
• Should I spend my mana on E.Reprise or should I waste mana by overcapping?
• Should I spend my mana on E.Moulinet during this trash pull or should I save it for E.Riposte...Scorch?

If we're going to answer these questions, then we need to take into account not only the mana that we're spending but also the mana we're not generating because we've given up a certain period of spellcasting in order to perform the spender. Spending X mana doesn't delay the next spender by X mana. Because all the spenders are on the GCD, they meaningfully delay the next spender by the mana that would have been generated if you'd used generator actions instead. Therefore, we have to take into account the foregone mana as an additional cost of using the spender. Only then can we measure one spender against another. Then, once we know what each spender is worth, we can set our expected potency-per-mana based on the spender with the highest return. Then we can assign value to the overcap mana in your scenario and arrive at the conclusion that it's obviously a damage loss to continue to cast spells if you're at 199 and you have the opportunity to perform a complete combo.

[Leidiriv]

That scenario isn't worth considering until we've already assigned value to a point of mana. What we are interested in are question like

• Should I spend my mana on E.Reprise or should I spend it on E.Riposte...Scorch?
• Should I spend my mana on E.Reprise or should I waste mana by overcapping?
• Should I spend my mana on E.Moulinet during this trash pull or should I save it for E.Riposte...Scorch?

Okay, so let's consider those situations.

1) E. Reprise has a potency of 220, with a potency per second of 100. As the melee combo has a raw mana cost of 125, let's compare 6 Reprises to the melee combo.
Melee combo: 2280 potency, 10.2 seconds, "effective cost" of 158-164
Reprise: 1320, 13.2s, "effective cost" of 160-169

2) To determine that, let's compare the potency per second of all the GCD actions we can take.
Reprise: 220 pot / 2.2s = 100 pps
Jolt + Dualcast: 560 pot / 5s = 112 pps
Fire/Stone + Dualcast: 590 pot / 5s = 118 pps
Moulinet: 200 pot / 1.5s = 133.33333 pps
Melee combo: 2280 pot / 10.2s = 223.52941 pps

From there we can compare the gain each action has over normal spellcasting and attain a "potency per mana" value.
Moulinet vs verproc: Moul pps x 1.5 - verproc pps x1.5 = 22.999 / 40 (mana spent) = 0.574 ppm
Moul vs Jolt/Dual: same formula as above but plug the Jolt/Dual pps from above = 31.999 / 40 = 0.799 ppm
Reprise vs verproc: -39.6 / 20 = -1.98 ppm
Reprise vs Jolt/Dual: -26.4 / 20 = -1.32 ppm
Melee combo vs verprocs: 1076.3999 / 125 = 8.6111 ppm
Melee combo vs Jolt/Dual: 1137.5999 / 125 = 9.1007 ppm

As such, in all scenarios it is a gain to overcap mana instead of spending it on Reprise, even before adding the mana opportunity cost in. It's even a gain over Reprise to sit around for up to 2.5s without casting anything at all.

3) The ppm of Moulinet only goes up for each target you fight, whereas the ppm of the melee combo just goes down faster and faster. Let's compare at 3 targets and at 4 targets, just to get a bit of data that way.

Moul vs VerAoE 3 target: 600 / 1.5 - 1020 / 5 = 400 - 204 = 196 x 1.5 = 294 / 40 = 7.35 ppm
Melee vs VerAoE 3 target: 223.52941 - 204 = 19.52941 x 10.2 = 199.19998 / 125 = 1.59359 ppm
Moul vs VerAoE 4 target: 800 / 1.5 - 1360 / 5 = 533.33333 - 272 = 261.33333 x 1.5 = 391.99999 / 40 = 9.79999 ppm
Melee vs VerAoE 4 target: 223.52941 - 272 = -48.47059 x 10.2 = 494.40001 / 125 = -3.9552 ppm

With these numbers, you can see that the melee combo is a gain over regular spellcasting even at 3 targets, but a massive deficit when compared against Moulinet. Once you reach 4 targets, it's not even worth looking at the part of your hotbar with the melee combo on it.

[Rongway]

Then I suppose by extension of that logic, wouldn't the extra 3 be cancelled out by virtue of the Verfinisher ideally proccing a Ready ability, which is 3 Mana more than Jolt?

Yes, you're mostly right about that. It gets more complicated because it's possible to go into the combo with both Verfastspells Ready, in which case the extra mana and damage that those Verfastspells would do over Jolt is already accounted for by the spells cast leading up to the combo; or it might be appropriate to use the "wrong" Verfinisher even without Acceleration, in which case there's only a 20% chance to generate those extra 3 mana and 20 potency. So the expected value will be less than 3 mana and 20 potency.

2) To determine that, let's compare the potency per second of all the GCD actions we can take.
Jolt + Dualcast: 560 pot / 5s = 112 pps
Fire/Stone + Dualcast: 590 pot / 5s = 118 pps

These are really the same case. The expected values of Jolt+Verslowspell are 250 potency and 6 mana for Jolt II, 310 potency and 11 mana for the Verslowspell, and 50% of the (20 potency and 3 mana) difference for the proc.

Jolt+Verslowspell: 570 potency, 5s, +18.5 mana

[the rest of the post]

I'm not sure what we're arguing about here? I never said anything that contradicts your recommendations about which spenders to use and when. All I said was the opportunity costs of not casting should be taken into account.

[Leidiriv]

I'm not sure what we're arguing about here? I never said anything that contradicts your recommendations about which spenders to use and when. All I said was the opportunity costs of not casting should be taken into account.

I'm arguing that time spent on a melee combo can't really be compared to time that could have been spent building up to a melee combo, because the entire point of building mana is for the melee combo.