Scorpion Venom vs Frost Spear

[Coprolith]

:D

Fun isn't it, to see an innocent little remark evolve into a heavy theoretical discussion?

/hugs

[Bryc]

All I wanted to point out was that Scorp was 8 ticks not 7 making it pretty damn effeicent. hehehehe

rofl

Your first mistake was using a decimal point. Decimal points are to geeks what Mad Dog is to alchoholics. When you typed "3.8", I saw "3(OMG PLEASE ANALYZE ME)8".


Coprolith,

I gotcha, good catch. My general equation should be:

effective damage per mana = (spell_damage - (lost_melee_damage))/mana

plugging in your formula:

lost_melee_damage = (casting_time*melee_dps)*(1 - 0.5/(casting_time/hasted_delay))

or

lost_melee_damage = melee_dps*(casting_time - 0.5*hasted_delay)

so, to plagiarize you:

effective damage per mana = (spell_damage - (melee_dps*(casting_time - 0.5*hasted_delay)))/mana

for a given mana cost, spells are

more efficient as they do more damage (duh)
more efficient the faster they cast
less efficient as as your melee damage increases
less efficient as your weapons get faster (surprising)

Which would cover the whole spectrum of slow 2hs to fast 1hs. I think =).

[Chubaka]

Do people back out to cast?  I just stay in melee range and channel it.  I get alot of procs imediatly after a spell lands.

[Coprolith]

I only back out when there's multiple unslowed hammering on me. I let the pet take that damage while i slow the mobs. Once slowed, i step back in. I never back out just to nuke or dot, the loss of melee damage that would cause defeats the whole purpose of nuking.
More procs directly after casting i believe are a figment of your imagination.

Which would cover the whole spectrum of slow 2hs to fast 1hs. I think =).

Krek, as long as casting_time > hasted_delay. I did this whole analysis when i was making my dps calculator, to see if i could incorporate combat casting into it. In the end I decided against it because it was too much work to implement for too little gain, and it would have meant 2 more user inputs which would only cause confusion. Nice to see my theoretical murmurings still coming in handy at a later time tho  :D

/hugs

[Chubaka]

More procs directly after casting i believe are a figment of your imagination.

Definitly not my imagination.  I think it is because I am engaged in combat and of course I don't turn off the attack to cast.  The "proc timer" (not sure what else you could call it)  decides that i need to proc but since I am casting, it waits until I swing after my spell lands.

I cast 2 dots on every mob and i would say on 5% of my casts I get a proc as soon as the spell lands.

[Coprolith]

While your proc rate is constant over time, the 'proc timer' is still tied to your combat rounds, not time. There's no clock ticking that says 'time's up, i must proc now' or ''i've just procced, tick tock tick tock'. You just get a chance to proc per combat, which itself is calculated in such a way that the average number of procs over time is independent of delay. If you lower your number of combat rounds per minute by chain casting you lower your proc rate. There's just no way to implement a code that calculates your chance to proc in combat round on a timed basis.

Trust me, in game perceptions are about as reliable as a politician's promises during election year. The random number generator is playing tricks on your brains. If you believe its not your imagination then i say parse it out and prove me wrong. If what you say is true you could make a macro that chain casts a spell with long casting time (IoS for example) and force almost every swing in between casts to proc. This should be easy to parse out against the Katta banker.

/hugs

[Tastian]

Too funny, chain casting IoS is exactly how I tested this awhile back.  I even went so far as to simply cast sha's + nuke + nuke every one minute and then totaled procs verse total procs well simply melee'n.

[Chubaka]

Too funny, chain casting IoS is exactly how I tested this awhile back.  I even went so far as to simply cast sha's + nuke + nuke every one minute and then totaled procs verse total procs well simply melee'n.

And what was your outcome?


Hrmmm Corp....all you say sounds good. hehehe.  To test this though I think I need to chain cast a clicky item cause mana would never hold out for a good parse.  But I also need a second of delay for a swing inbetween casts?

I am curious about the average over time still though.  If I was to aug that 45dmg 150dly hammer from da giants in burning woods and used no haste items or haste spells, this thing would have to proc every second swing on average?

I think I will burn 40 adventure points on that 20dd fire aug to test it out. heheh seems like fun.

[Coprolith]

Tastian wrote:
Too funny, chain casting IoS is exactly how I tested this awhile back. I even went so far as to simply cast sha's + nuke + nuke every one minute and then totaled procs verse total procs well simply melee'n.


And what was your outcome?

Yes, inquiring minds would like to know  :D

Hrmmm Corp....all you say sounds good. hehehe. To test this though I think I need to chain cast a clicky item cause mana would never hold out for a good parse. But I also need a second of delay for a swing inbetween casts?

Well statistics on proc rate are sucky, no doubt about it, but the effect of chaining IoS should be so profound it wouldnt take long to see its effects. I think you can chain it for 4-5 minute in a row, med back to full for another few minute, rinse and repeat. All in all i dont think it should take longer then 30 mins before the effects become clear. There's no real need to parse the baseline (= proc rate without casting), that's pretty much a given.

I am curious about the average over time still though. If I was to aug that 45dmg 150dly hammer from da giants in burning woods and used no haste items or haste spells, this thing would have to proc every second swing on average?

Wasn't that a 2HS? Seems like ages since ive killed giants in FM. Regardless, you're correct. With a really slow 2HB, augmented, no haste (better yet slow yourself) you'll pretty much get a proc in almost every combat round.

/hugs

[Eatbugs]

With a really slow 2HB, augmented, no haste (better yet slow yourself) you'll pretty much get a proc in almost every combat round.

That's the trick the Warriors who were AEing in Fire used with Earthshaker (70 delay 2HS that procs Earthquake) - they had themselves slowed ahead of time in duels with Shaman, then pulled everything in the first field and hit Rampage (Warrior AA that attacks all targets in range) and had procs on nearly all hits.

Of course, they died immediately afterward to the few mobs left alive, but it made for some really amusing logs.

[Dakat]

When calculating dps of a spell shouldn't you use:

From the start of casting of the spell until the landing of the spell on the mob from the second cast.

For instance.

1. Your casting SV on a mob. You start to cast (clock starts). Without affliction haste item, casting time is 4.1 sec. Spell duration is from initial hit and 8 tics. Then recast spell again for another 4.1 sec, spell lands (clock stops).

2. Your casting FS on a mob. You start to cast (clock starts). Without affliction haste item, casting time is 3.2 sec. Spell is instant duration. Then wait 30 seconds before spell can be recasted. Then recast spell again for another 3.2 seconds, spell lands (clock stops).

From the time you start to cast your first spell until the landing of your second spell, the total damage that you do from that spell is during this (window) is your total damage from the spell. You might be doing 1278 dmg (110 initial and 146 per tic for 8 tics) for SV without dmg bonus focus item. Calculate 4.1 sec cast time, initial land spell for 110, 8 tics at 146 per tic, each tic is 6 seconds apart, from the initial hit to the last tic is 54 seconds, recast spell takes another 4.1 seconds.

So your doing 1278 damage using Scorpion Venom in 62.2 seconds or 1 min 2.2 seconds.

For Frost Spear, 675 damage, 3.2 sec casting time, landing instant, 30 second recast time, 3.2 seconds to recast spell.  Is 36.4 seconds. Putting the 2 in comparison this way they both are doing roughly the same amout of damage in the same amount of time.

This is just my thoughts on the subject anyways. I could be totally off base in my assumptions.

[Dakat]

While I was on a mind break, I started thinking too. Yeah watch out, its flying now...

Calculating melee dps.  Shouldn't you calculate the total number of swings taken during a given period of time.  With your misses, dodge, block, riposites, hits included. Not every time are you going to perform the exact same way on each mob.  

Someone mentioned something about hybrids.  We are hybrids, in the fact that we can not melee and cast a spell at the same time. If your casting a spell, or a proc goes off, your not meleeing at all. You might get a hit between, but its probably off a riposite and not an actual swing.

It's late here and I think I burnt my brain up.  someone enlighten me here.

[Dakat]

Going back thru some of the posts again. Missed some things from previous posts.

You must have some form of improved damage on your spells. I get 146 per tic on SV withouth any thing added.

There can be so many variables when calculating dps for these spells. Not every person will be the same.

Say for example. I'm wearing a mutititude of focus affects that enhance every aspect of my damage spells. Such as Affliction Haste, Affliction Efficiency, Burning affliction just to name a few. I now have got shorter casting time, spells last longer, spells doing more damage per tic.

[Tastian]

Ok let's see...

-  No bst has a cast time of  3.2 on nukes or 4.1 on any dot other than like plague.  We get hybrid spell haste.  At 65 nukes are 1.9(ish) casting time.  Scorp is around 2.1 (ish again I just woke up and I'm already late lol).  

-  You don't factor in hits/missing/etc because in any given 1 round anything can happen. Rather you look at the average dps and average results over time.  Sure in one cast you might get a resist, might get a crit for 1485+.  You also might have quad'd for max every hit and crit every hit.  Look at the average damage over time and work around those numbers.

-  Hybrids can't melee well casting, however, hybrids were "tweaked" awhile back so that spell casting wouldn't recent the melee timer.  Originally if you cast a spell when it finished you restarted your melee round.  Now when you finish casting you attack right after in most cases (depending on weapon delay and spell being cast).

-  Casting spells lowers proc rate

-  I get "some" procs right after a spell finishes, but I also get some spells that "restart" casting because I get a proc just as my spell is about to start casting.  Basically I click the button to cast the nuke, the spell gems flicker like it was a fizzle, the proc fires, then instantly after that the casting bar appears and the spell starts going.  

-  If you had a MH weapon with innate proc and gave it another aug'd proc, as well as an off-hand weapon with a proc and aug'd it with a proc and had WA5 then you would expect (in most cases)  ~9 procs a minute or one proc every 6-7 seconds.  If you start melee'n something and simply cast Spiritual vigor (5sec casting time) you simply won't proc after EVERY cast finishes.  Take a few minutes of casting SV and watch that you don't proc after every swing.  Also if having attack on kept the timer going then people could kite mobs with attack on then dart by and weapon would proc right away because you'd been in "attack mode" but had not had a chance to proc yet.  

-  I'm late, talk to ya's sunday.  lol

[Coprolith]

It's late here and I think I burnt my brain up. someone enlighten me here.

Let me first give you short version of this reply because i fear the long version could cause permanent brain damage  :D

The short of it is that spell dps or efficiency of a spell is  very situational and the method of calculation should be tailored to that situation.


The long version:

When calculating dps of a spell shouldn't you use:

From the start of casting of the spell until the landing of the spell on the mob from the second cast.

Yes thats one way of defining it. The dps of a spell is a tricky one. First of all there's more then one way to define dps. Secondly you have to differentiate between DoTs and DDs since they deliver their damage in different ways. E.g.:
1. max potential spell dps, or burst dps, is equal to spell_damage/(casting_time+recast_time) for DDs and total_spell_damage/(casting_time+DoT_duration) for DoTs.
2. max sustainable spell dps is equal to spell_damage * mana_regeneration_rate/mana_cost (provided mana_cost/mana_regeneration_rate > cast_time+recast_time or > cast_time+DoT_duration)
3. actual sustained spell dps is taken over a long period of time by summing up all the damage done by that particular spell divided by the total time. This is the closest to the definition you're giving since its averages out to spell_damage/average_time_inbetween_casts.
4. "real" spell dps is the total damage done by that spell divided by the time it took to kill that specific mob from the moment it was first engaged. You could then average out this dps over many fights.

#2 and #3 are very similar btw. For max sustainable dps you can read mana_cost/mana_regeneration_rate as the shortest possible average_time_inbetween casts. #4 differs from #2 and #3 in that the latter 2 assume there's no downtime whatsoever.  #4 is the 'real' spell dps, that is it applies to normal gameplay and its sort of a mix between burst and sustained spell dps. Its very situational tho which is why its so hard to compare spell efficiencies in the first place.
#1 and #2 are easy to calculate and are therefore better suited for a general comparison (#3 and #4 you have to measure, not calculate). Unfortunately in the case of spell dps, or spell efficiency, the general comparison doesnt have as much bearing on real gameplay as is the case with melee dps because spell damage isnt nearly as 'smooth' as melee damage.

Calculating melee dps. Shouldn't you calculate the total number of swings taken during a given period of time. With your misses, dodge, block, riposites, hits included. Not every time are you going to perform the exact same way on each mob.

Again it depends on how you define dps. You can (theoretically) average out your melee dps over many different mobs. A sort of 'lifetime' average provided you never upgrade again or get dps boosting AAs. Every time you do upgrade, you reset the average and the clock for the 'lifetime' average starts ticking again. Or you could look at sustained melee dps for a partiular situation.
In the discussion we had before in this thread on how you should correct spell efficiency for loss of melee dps you can use either, tho generally you won't know your 'lifetime' average. However there's little harm in using your sustained dps measured for a particular situation and apply that to a different situation. Its only a correction after all, and a small error in the correction factor can be considered a second order effect.
For the purpose of a general weapon comparison you'd want to look at the lifetime average dps of that weapon. Then the next step would be to look a little deeper and compare weapons for a particular situation, in which case you should use the sustained dps for that situation.

I don't know if this cleared things up for you or if i just permanently put your lights out  :wink:

/hugs