Note that cleric damage isn't accurate, because I have no good numbers for cleric melee damage with the new proc hammers.  These graphs only show cleric nuke damage.

Also note that shaman DoT's are used, and shaman DoT's are taunting.  It takes a high quality crazy glue warrior to prevent the shaman from becoming the tank when using these DoT's, so this damage isn't particularly realistic.  That said, it looks like (from analyzing spdat modifications) shaman DoT's in PoP will be significantly less taunting than current versions.  In particular they do not have an initial hit, nor do they have very many poison/disease counters, both of which are major factors in DoT taunt.  Shamans may very well be able to pump out high damage per second in PoP without gaining aggro, but we'll have to see.

Rain spells are graphed, but since they can't be used in every situation, they're simulated seperately from non-rain damage.  Not everyone has their ancient nuke/DoT, so that too is simulated seperately.

Update 10/19/02 Clarifications: All DD damage, by druids and clerics, is increased by 10% from focus items and 7% from crits (Spell Casting Fury 3).  This was originally part of the sims, but wasn't mentioned anywhere.  Oddly enough, the supposed lack of the 7% damage increase from crits was one of the primary shaman objections, even though it would be trivial to look at the graphs and mentally adjust them up 7% (don't do it though, as I said, it's already included). The actual calculation is to multiply by 1.1, then multiply by 1.07...which actually results in a slightly higher number than what's currently live, because the way these things are calculated was changed.  But it's pretty close. These damage calculations lack these features, that I can think of off the top of my head: Melee haste (shaman) Pet (shaman primarily, druids to a lesser extent, cleric pet takes mana and is more of a nuke/DoT) Druid epic (I should really add this, but I forgot to) Damage shields (hard to quantify, but it definitely adds damage to druids) Cleric melee (I have no numbers for this) And probably some other things.  Shamans also objected about the missing items above, except that oddly enough, no shaman objected that their pet was not included.  Shaman pets, from the last time I parsed one, do about 10 dps.  However, due perhaps to leash laws, shamans rarely use them. Additions 10/19/02 I've added another sim to the list, the short damage (fights less than 1 minute) sim.  This cuts out any DoT's longer than 60 seconds, i.e., Pox.  When doing this, I noticed that the sim shaman was using up all his mana on Ice Blast instead of keeping the more effecient DoT's active, so I changed the script so he would only Ice Blast only if he had excessive mana, and damage/sec went up.  That's the problem with simulated shaman, they aren't very bright.  I said simulated shaman!  No flames please. I updated the original sim the same way (smarter shaman). These sims don't count resists, but neither do they count resist debuffs. Corrections 10/21/02 Sim shaman was using his epic in the fast damage simulation, even though he shouldn't have.  I took it away, and sent him to his sim room. Additions 10/21/02 Added a 'nukes only' simulation.  As the name suggests, this compares the classes in situations where the only damage spells they use are nukes (Moonfire, Ice Strike, and Judgement). Comments 10/30/02 The DoT simulations assume that the shaman casts a new DoT every time the old one wears off, but in real situations, it may be possible to cast the same DoT again early (if a second mob joins the fight), or it may not be possible to cast it when it expires (because the mob is near death, and no new mob has arrived). What this means is that the first few minutes show maximum damage/sec, but only if fighting one target the whole time.  In some situations it won't be accurate.  The 20 minute value is accurate for one mob or multiple mobs, since the real priest can cast just as many DoT's as the simulated priest - he just isn't likely to cast them in exactly the same order, at exactly the same times.  The real priest can also avoid the aggro generated by the simulated priest, by not casting all the DoT's on one target. Depending on the situation, it may not be possible to use all the DoT's the simulation shows, in which case some mix between the all-DoT simulation and the all-nuking simulation is closer to reality. For example, shaman DoT's can be particularly taunting, so in a group where there is no strong tank, a shaman trying to do primarily damage will likely cast one or zero DoT's per mob, and mostly nuke.  In this situation, it's the nuking-only sim that makes sense. Changes 10/31/02 Congratulatate sim druid, today he completed his epic! Sim druid now uses his epic in all damage sims except nuke-only.  In the one minute fight sim, the epic effect lasts only one minute.

Commentary

Note that 100% fighting is not possible in the real world (I enjoy using the phrase 'real world' to describe a game - I don't know why).  In practice, there is time lost while mobs are being pulled, during buffing, etc.  That time translates into higher real dps for casters relative to melees.  At the same time, the above graphs assume zero resists, and in practice, resists happen.  Thus it's hard to use any graphs showing caster dps in a meaningful way when comparing to melee damage.

In situations where there is some downtime, the shaman advantage grows, because shaman mana regeneration is directly proportional to how much time a shaman has to cast cannibalize IV and torpor.  It's the conflict between the cast time of mana regeneration spells and the cast time of damage/healing spells that led me to create simulations instead of attempting to calculate mana regen and multiply it by effeciency.  The actual values in are necessarily different, because the more heals/DoT's casted, the less time there is to cannibalize.