Vontre's Mage DPS Spreadsheet - Page 3

Micah · 04/20/07, 7:08 PM

Hmmm.... Anyone know where I can buy some +Reading gems IRL?

◊ Vontre · 04/20/07, 7:31 PM

Originally Posted by Micah

Hmmm.... Anyone know where I can buy some +Reading gems IRL?

http://www.lenscrafters.com/

Ghalt · 04/23/07, 11:58 AM

I have it up all the time. I play with different builds, gear upgrades, etc.

I'm trying to tweak my build to be the best for the raid.

One thing that I'd love to see is the ability to theorycraft various spell sequences that don't involve Arcane Blast.

For example, what would the DPS and Time to OOM be for Fireball x2, Scorch, Fireblast, repeat.

Ghalt · 04/24/07, 4:36 PM

From what I'm reading, a player's own Ignites will roll/stack.

Can that be accounted for in your spreadsheet? That seems like a serious piece of work to include that, but would make a big difference in dps based on crit gear on long fights, I would think.

I know you have a life, and aren't at my behest to keep updating the spreadsheet on my whims. I wish it were so, but alas...

But seriously, thanks for your hard work on this, and I'd love to hear your thoughts.

◊ Kalman · 04/24/07, 5:18 PM

Rolling ignites, if they're personal, can probably be modeled, but it's not worth putting in the effort until it's confirmed that the mechanic is making a return.

Zene · 04/25/07, 1:26 PM

Hi Vontre,

Great job on the excel spreadsheet.

You might have covered this already, but I was curious how you modelled +spell hit rating in your DPS calculations. There seems to be a couple of schools of thought the 1 roll system versus the 2 roll system.

Thanks,

Z

Vendethiel · 04/25/07, 5:29 PM

You are able to choose, wheter you want to apply the 1-roll-system or the 2-roll-system... On the Main Sheet, some way down in "Misc." area there ist an option you can set: "Two-Roll System***" Either set it to yes or no depending if you want to use it or not...

◊ Vontre · 04/25/07, 5:57 PM

Originally Posted by Ghalt

From what I'm reading, a player's own Ignites will roll/stack.

Can that be accounted for in your spreadsheet? That seems like a serious piece of work to include that, but would make a big difference in dps based on crit gear on long fights, I would think.

I know you have a life, and aren't at my behest to keep updating the spreadsheet on my whims. I wish it were so, but alas...

But seriously, thanks for your hard work on this, and I'd love to hear your thoughts.

=D thanks. I do most of this when I'm bored at work, actually.

Rolling ignites is an interesting concept to model. I would be excited about their return if they were compensated for by a reduction in direct or critical damage from fire to compensate... yeah. I like the idea, I miss the old fire blast after an ignite proc, but I'm not too excited about fire specs further pushing ahead of ice/arcane on damage charts.

Here's what I'm thinking. Chance to roll an ignite is the same, basically, as chance to crit two spells in a row. Or in the case of scorch not in a row, but we're still talking about a crit on two spells. The chance to crit two spells can be calculated easily enough, so I suppose that could serve as a decent "ignite proc chance" and count as added average damage. If you have say 7% chance to roll an ignite, 7% of the added damage is applied to the average.Which would probably end up 20-30 additional damage overall to the average. Which would be roughly 1-2% increase for overall dps.

I could add more tables for chance to proc a third, fourth, and fifth ignite, which increase in power but decrease in likelihood.

Liebestod · 04/25/07, 8:12 PM

Hey there, I was directed to this sheet when I was looking into the possibility of creating something similar to this myself. Nice to see others have done most of the work for me, though. I'll probably try to peek into the code in the future, though, to see what's going on.

A couple things I'd like to suggest, though:

1) Having more flags for specific item effects would be nice, eg. +dmg trinkets, Nexus-Horn, Mark of Defiance, Quagmirran's Eye, and more set bonuses. I'm running a crit-heavy Arcane/Fire build at the moment which isn't doing too hot on the DPS calculations, but is good on efficiency without mad raid buffs.. I'm wondering how it would be affected by being able to add these kinds of considerations. I also wonder if the current spreadsheet can factor in things like AP/PoM/Pyro every 3 minutes into DPS.. kinda doubt it, but these considerations are relevant.. maybe I'm just trying to console myself though.

2) Having a feature that calculates the marginal DPS/longevity bonus of an additional point of various stats would be pretty useful and easy to implement - right now it can effectively be done by just adding a 1-unit upgrade of the various stat, but it's a bit cumbersome, especially when one would like to compare multiple stats at once. If the marginal values could be weighted based on their slotmod value, then very useful itemization advice could be given: "The highest return given your current data would be to invest in _________."

3) Spell penetration should probably be included somewhere, but it would only be useful if target resistances could be imputed. That data isn't always forthcoming (at best one could use long-term resist rate data to try to guess the values for various targets), but... it's a huge omission atm.

4) As was already mentioned, more rotations could be added. I generally use a Scorch/Fireblast rotation, for example, which can't be modeled atm. I realize that factoring FB in would be difficult because its cooldown could be altered by talents, but at least having various rotations that use 0/3 Imp FB and 3/3 Imp FB would be useful. Also, I'm curious as to how effective a Scorch/FB/AB rotation would be... I know this is even more difficult to model, but it'd be nice to see the more common rotations thrown in.

Overall.. great work so far! I'm not very familiar with spreadsheet coding, but if possible I might try to code in some data that could be implemented in a modular fashion - maybe more of the flags I mentioned earlier.

◊ Vontre · 04/25/07, 8:44 PM

Thank you, though in the case of those trinkets modeling them when they are often used in conjuction with player abilities like arcane power is fairly difficult.

Modeling additional rotations is not hard, it's cumbersome from a user-interface standpoint. I'll look into it.

Spell penetration is pretty irrelevant, until this changes I'll be ignoring it completely.

Regarding #2: So you've been using the upgrade comparison module to compare stats, but you're saying you'd like a module to compare each stat at once? Interesting... this is the one I'm most likely to actually implement =)

Liebestod · 04/25/07, 8:58 PM

Regarding #2: So you've been using the upgrade comparison module to compare stats, but you're saying you'd like a module to compare each stat at once? Interesting... this is the one I'm most likely to actually implement =)

I originally became interested in modeling these things because I found myself wondering what types of stats I should be focusing on given my current build - knowing that the more crit I had, the more additional +dmg and +hit would help my DPS, I was wondering if I had passed the threshold at the returns to these other stats would be greater than stacking more crit. According to the spreadsheet the answer is "yes", but like I said, it's a bit cumbersome to make this comparison given the current interface. Also being able to adjust the comparison based on slotmod (eg. compare 1.2 +dmg to 1 crit rating, since crit is more expensive to itemize) may be useful.

Modeling additional rotations is not hard, it's cumbersome from a user-interface standpoint. I'll look into it.

Well, more choices won't confuse us that much. :P I'd suspect that mixing in fireblasts would raise the DPS of a Scorch build by 10% or so, as well as lowering its efficiency. It'd just be nice to be able to be sure.

Thank you, though in the case of those trinkets modeling them when they are often used in conjuction with player abilities like arcane power is fairly difficult.

Yea, I imagine that there's a margin at which it's no longer useful. But trinkets can still be modeled usefully in a very simplistic manner - eg. if a trinket with a 2-minute CD gives +150 damage for 15 seconds, this averages out to 150/8 +dmg in the long run, and that conversion could be factored in to calculations. Some trinkets like the Lightning Capacitor, Mark of Defiance, etc. probably have a rather significant effect on DPS or efficiency and deserve to be factored into consideration somehow, if only in an over-simplistic manner.

Oh yea, and I should've added a 5th suggestion to my previous post: Threat per second. Should be relatively simple to include, and... seems useful?

[Edit]

Also, one other comment... the Vampiric Touch flag seems to based around a static amount of Shadow Priest DPS. It would be nice if it were adjustable, since I imagine that it's default level may be rather high compared to what I'm used to..

Roywyn · 04/25/07, 9:19 PM

Originally Posted by Ghalt

From what I'm reading, a player's own Ignites will roll/stack.

Tried it a few days ago on the PTR, and they didn't seem to stack.
1k crit, 200 tick, 1k crit, 300 tick, 300 tick, just like on the live servers.

I did get some strange result once when I crit just when the second tick of the previous ignite ticked, leading to 2 new ticks of some pretty odd damage, higher than it should be. I don't remember the numbers sadly. I still think of it as a lag-related bug though, the same that eats my ignite damage on Fireball + Fire Blast combos.

And yes, ignites that do not reset tick timers are a pain to model.

◊ Vontre · 04/26/07, 4:11 PM

New version is out - v2.5. You can get it from the same place as usual (see my signature) and you can delete the mage_dps_gear.xls copy that's been floating around.

I finally got around to merging the gear selector with the main sheet and actually making it useable. Big thanks to Kalman (Laplace) for writing up all that data!

The way it outputs your stats is a little different now with the gear selector in place. There is a yellow section in the 2nd column that will show your stats, depending on whether you opt for gear selector or manual stats. There is also a big, friendly yellow option right at the option that lets you toggle between the two.

Gear selector is now the second page, instead of manual talent selector.

I'll be looking at some more of the feature suggestions in the future. Thanks for your support everyone!

Leto · 04/26/07, 4:56 PM

Might want to add a note about the manually editing talents option.

Also, the gear selection doesn't seem to update school specific dmg, and manually setting school specific dmg adds to both the manual section's stats and the gear specified stats.

I'll let you know if I see anything else.

Wynna · 04/26/07, 8:02 PM

First off, thanks a lot for this effort in making the mage spreadsheet look more like the rogue one, I've alllllways wanted that so I could make multiple gear swaps and see the results.

Now to the feedback, first thing I noticed was the lack of items to choose from, I have like 4 items that are not in the spreadsheet, Frostfire Robe, Ring of the Eternal Flame, Neltharion's Tear and Nethekurse's Rod of Torment, I tried adding those myself manually but it didn't work out, they wouldn't show on the drop downs, nor did my leet +5 spirit enchant to boots (hardcore pver lawl), it also doesn't seem to check to see if you have the right gems to give you the socket bonus and most of the trinkets don't seem to be modeled but I'm sure you'll do that eventually. I think that's all, if I notice anything else I'll let you know.

ps: /w vont go get minti vont.. DO EET

◊ Vontre · 04/26/07, 9:40 PM

Originally Posted by Wynna

ps: /w vont go get minti vont.. DO EET

Minti was gotten, I pwned him. Ice mage > combat rogue. =p

On topic. I actually did another revision today and sent it to Kalman. It's 2.5.1 and it has the spell damage values fixed for gear selector. It also features a brand new cycle builder. Lets you custom build cycles of... I think 12 or more spells and add them to the main sheet. I think it's neat... hope you guys like it.

Wait on Kalman for new upload. =D

◊ Vontre · 04/27/07, 4:47 PM

The new spreadsheet is uploaded, you can get it now. I put in a few premade cycles like Fireballx2 Fire Blast, or Scorchx3 Fire Blast.

Kavan · 04/27/07, 11:45 PM

Now that it's possible to enter gear and with all the AB cycles available I decided to make a comparison what results I get compared with my spreadsheet. This is for a pure arcane build and I was surprised that ABx3+AM+Scorch was preferred compared to ABx3+AMx2 that my spreadsheet prefers. Now granted the difference is not that big, about 870 vs 890 dps, but it was enough to pick my interest in trying to dig down where the actual difference materializes. I'm not sure I found it, but the best explanation I can make is that the benefit of AMx2 cycle comes from higher proc rate ability when used with Shiffar's Nexus-Horn, The Lightning Capacitor and similar gear. I'll stay tuned and am looking forward to make some more comparisons when some of the proc abilities are implemented.

◊ Vontre · 04/28/07, 2:58 AM

Originally Posted by Kavan

Now that it's possible to enter gear and with all the AB cycles available I decided to make a comparison what results I get compared with my spreadsheet. This is for a pure arcane build and I was surprised that ABx3+AM+Scorch was preferred compared to ABx3+AMx2 that my spreadsheet prefers. Now granted the difference is not that big, about 870 vs 890 dps, but it was enough to pick my interest in trying to dig down where the actual difference materializes. I'm not sure I found it, but the best explanation I can make is that the benefit of AMx2 cycle comes from higher proc rate ability when used with Shiffar's Nexus-Horn, The Lightning Capacitor and similar gear. I'll stay tuned and am looking forward to make some more comparisons when some of the proc abilities are implemented.

You do not benefit from debuff overlap if you use 2 waves of arcane missiles.

Kavan · 04/28/07, 3:32 AM

Originally Posted by Vontre

You do not benefit from debuff overlap if you use 2 waves of arcane missiles.

I'm aware of that, but AMx2 still comes on top, the difference in dps of AM and Scorch is that big. Dps of AM is actually higher than ABx3 ramp, I only use the cycle to increase mana efficiency. Of course on short fights AB spam is the only option.

◊ Vontre · 04/28/07, 3:41 AM

Originally Posted by Kavan

I'm aware of that, but AMx2 still comes on top, the difference in dps of AM and Scorch is that big.

That doesn't really make sense. The overlap is such a powerful mechanic, as soon as I started modeling it every non-overlap mechanic pretty much dropped off the radar. I can't be sure how you got that result without seeing the calculation. Keep in mind my cycle builder is not just for dps, it meets mana consumption restrictions as well.

Kavan · 04/28/07, 4:00 AM

Originally Posted by Vontre

That doesn't really make sense. The overlap is such a powerful mechanic, as soon as I started modeling it every non-overlap mechanic pretty much dropped off the radar. I can't be sure how you got that result without seeing the calculation. Keep in mind my cycle builder is not just for dps, it meets mana consumption restrictions as well.

I know how it works, I've been digging through it

. My spreadsheet optimizes the spell mix as a linear program. So it's similar in that you specify all parameters of the fight like duration, resistances and so on. Except I don't match mps with going oom manualy, but just let the LP solve the problem for me. But that aside in my calculations the ABx3+AMx2 vs ABx3+AM+Scorch I get the first one dominating the other in both dps in dpm, comparing in a similar way to how you've done your cycle builder.

Actually now that I've done some more tests it seems that the more important factor is different miss rate between arcane and fire. I've disabled the trinkets and still get AMx2 ahead, but AM+Scorch is ahead if I don't have resists. This makes me wonder since you model this part as to where the discrepancy is. I'll try to figure out what's going on and let you know what I find.