[Hunter] The Survivor - Page 5

Cheeky · 05/30/07, 10:57 AM

Originally Posted by Elakuan

Yes...I've been putting together a "dream" set of marksman stuff and survivor stuff from kara, since I think that's where the biggest opportunity for taking advantage of this lies (where I am now, ssc and the eye it gets a bit more stablized).

Kind of starting to feel like a moot argument, but I'll keep tinkering on it.

I think my fear is that in doing so the argument will be reduced to "oh, well for THAT set of gear sure." I was trying to avoid too much specificity. Perhaps the better way to do it would be to look at the way blizzard itemizes as a whole, and do chalkboard comparisons that are applicable to all levels of hunters.

Hmmm.

One way to control for that is to find the best items for each role from a certain iLevel of gear. Karazhan is almost exclusively iLevel 115 gear, with the Prince dropping 125. We could then look at all mail and leather items in that range and mix and match. A spreadsheet is useful for quickly switching out 1 item and seeing how that effects all the derived calculations.

Unfortunately I do not track iLevels in my spreadheet, but WowHead and Thottbot do, so it wouldn't be too hard to cross-reference. (Or I could build the iLevel formula in and calculate gear on the fly.)

I don't think the argument is moot at all. As a DPS-whore I would like to know if I'm better off for both the raid and personally switching specs. I only went BM when I had proved to myself mathematically it was more DPS than my MM build.

Elakuan · 05/30/07, 10:57 AM

I just had a conversation with a friend about this, which helped me clarify my point a bit more. And its a simple point.

Blizzard builds mana tools around its casters. These casters typically have HUGE mana pools and expensive rotations.

Of any class/tree that uses mana...we survivors USE it most effectively. We cast a lot of spells without spending a lot of mana.

Fact is, water is designed around people who have to replenish twice our mana pool. A cleaner design would be water/potions that restore some set percentage of your mana pool...but they have given us a fixed coefficient.

Which is where the opportunity lies. By building a survival-mana-efficient machine...we can best take advantage of those tools, and if we choose sacrifice some of that startling longevity for some dps gear...making our tree viable.

"how viable" and "how sustainable" are interesting questions. But by not giving us percentage-based mana tools...blizzard has opened the door for us.

Elerion · 05/30/07, 11:00 AM

The solution to that is to do two or three sets of modelled numbers, one for each level of progression. Say one comparison for Blue, one for Kara/Heroic, one for SSC+. Make optimal sets from each of those levels for each of the specs, and present numbers. Allow others to make suggestions to optimize the sets or underlying figures. From this, a conclusion could be made along the lines of (example):

"In a theoretical vacuum, a blue hunter going survival and agi gear will outdamage a Marks hunter with balanced gear. The difference diminishes with Kara gear and is reversed when we get to SSC+ gear."

By having a framework ready, you could easily modify this to answer any "But what if..." comments that arise, and your conclusion will have some definite value.

Elakuan · 05/30/07, 11:00 AM

Originally Posted by Cheeky

I don't think the argument is moot at all. As a DPS-whore I would like to know if I'm better off for both the raid and personally switching specs. I only went BM when I had proved to myself mathematically it was more DPS than my MM build.

Well, I have little doubt of BM's superiority at this moment. Until my other thread on haste rating comes through and I can combine it into this one...I think the fact that blizzard has closed the gap on me (ranger-general's chest, for example) makes this less of an end-game issue.

But as time permits, I'll put something together.

Elerion · 05/30/07, 11:03 AM

Another thing that would be interesting is longevity calculations. Calculate how much negative mp5 an optimal survival cycle has compared to an optimal marks or BM cycle, and you would be able to calculate longevity. Compared with an analysis of the time cost of drinking and a dps analysis, this would allow you to make conclusions regarding the validity of each build as a function of fight duration.

If you could combine this with the analysis above, you would be able to plot the dps of each build and gear level in a dps/fight duration diagram, or you could plot certain combinations of the above in a single axis column diagram to simplify it. I imagine the value of something like that would be immense for quick reference.

Lactose · 05/30/07, 11:18 AM

Originally Posted by Elerion

Compared with an analysis of the time cost of drinking and a dps analysis, this would allow you to make conclusions regarding the validity of each build as a function of fight duration.

A cleaner way of putting this, in my opinion, would be something like this:

TotalDPS = (NormalDPS*[FightDuration-TimeDrinking])/FightDuration

TimeDrinking being whenever you have DPS breaks, forced (happening regardless of spec due to encounter, etc) + due to OOM (spec and gear based).
NormalDPS being DPS dealt while not OOM.

Elerion · 05/30/07, 11:32 AM

Actually you could remove TimeDrinking as a variable with some work.

Calculate the following:
When OOM, how many seconds do you have to spend drinking to facilitate one second of DPSing?

Make a ratio:
TimeDPSing/[TimeDPSing+TimeDrinking] = DPSratiopostdrink
with the results from above (say 1/[1+0.2]=0.833).

Now you could make a sum function where TotalDPS is your full dps for the duration you can sustain without drinking, and the full dps times the DPSratiopostdrink for the remaining duration.

This would yield a TotalDPS function that is only dependant on one variable (Fight duration) in the final iteration*.

This function wouldn't be able to incorporate forced breaks from fight mechanics, but it would be more straightforward for general use. And unless you are trying to choose the perfect spec for one specific encounter, generalized formulas seem more useful, not to mention easier to use for those that don't know how to choose the perfect values of variables like TimeDrinking.

* Obviously it would also be dependant on a lot of variables that help build the DPSratiopostdrink and NormalDPS variables. These would presumably be given from other functions in the spreadsheet however, so the only exogenous variable would be fight duration. In other words, once the rest of the spreadsheet is working, you could simply enter fight duration and get a TotalDPS number, or you could plot TotalDPS as a function of fight duration for a more general view.