That doesn't look right. You can get the exact average with recurrence on number of casts like this:

activeTime[n] := p * (min(n, D) + activeTime[n - C]) + (1-p) * activeTime[n - 1]

where p is proc chance, D duration and C cooldown in terms of number of casts. The graph looks like this:

Kavan: Ours are the same, yours is just smoothed. I haven't used a program capable of smoothng nicely since I was in school (used to use R) so I didn't bother. I'm also not sure smoothing it makes sense, since in practice you should be seeing an angles step function where step length is equal to cast time.

It's not the same. It's completely different shape, it's not that jagged at all. Here's a nonsmoothed version:

You have two different variables - one is tracking fight length, the other is tracking # of casts. They're going to look slightly different unless the data frequency for the fight length x-axis is 2.5 seconds. The former graph looks more spiky due to the dropoff after the spell is cast, then jumps up again when the next spell is cast.

Edit: Upon closer inspection it looks like the timing might be a bit off on Hotan's graph. I'd think that the first jump would be at the 2.5s mark.

With only one cast, you have 0% uptime. The first spell may proc it, but since there is no second spell the proc is used for no spells. Thus until 5second duration (2 casts) there is no possible uptime benefit.

Another point of difference is that my graph actually assumes 5seconds between ICD and proc in all situations. While I realize this is incorrect, it makes the math more understandable to the average reader, and saved me re-installing a more powerful statistical program I don't currently use. This simplicity doesn't change the data in a significant way.
The only difference in our data is the shape of the local min/max.
Your formula is taking into consideration the intra-ICD time is not static, but rather a geometric function of p.

Like I said before. Your graph is the same as mine, it is just smoothed at the local min/max (ie the time of proc and proc fade)

However I find the average 50% uptime with 1 casts flat out incorrect. The chance of the first spell proccing is 50%. If it procs the uptime for the first two spells is 1 cast. If it doesn't proc the uptime for the first two spells is 0 casts. 50%*1/2+50%*0/2=25%. Your graph very clearly shows this as 50%, which is flat out incorrect. You are missing the point that it doesn't matter at all if the last spell cast procs.

Also, I do find the fact that your probability never falls bellow 30% suspect. In a 55second fight (2.5sec cast time) you can expect one proc at 5seconds and a second proc at 55seconds, however the second proc produces no uptime. Therefore you have 15/55 uptime = 27.3% uptime. This is the expected uptime. I believe this is caused by the same issue as before, it doesn't matter what the last spell does since you won't use that uptime, if there is any.

To correct this last spell concern you should shift your uptime by one interval, while keeping the duration static.

Ok you're right, in my model I started counting time at the time of the first cast, all you need to do is change it to off by 1 like this:

activeTime[n] := p * (min(n-1, D) + activeTime[n - C]) + (1-p) * activeTime[n - 1]

which doesn't change the shape at all.



The fact that you're assuming exactly 5 seconds between ICD and proc in all situations is my point of contention. You can't assume that. Specially when you have procs with lower chance than 50% you would get completely wrong result. You would predict that it remains jagged through all time with fixed period while in reality it's a lot more smooth and oscillation keeps increaseing it's period and eventually disappears.

This is just not correct. You arrived at this conclusion by assuming it takes exactly 5 seconds to get the first proc. The chance that you'll get more than one proc (and have nonzero uptime) on 55 second fight is not 0. At the same time you forgot that there is a admittedly very tiny chance that you won't get even one proc. When you correctly take all possibilities into account you'll see that 15/55 is not expected. You might call it an approximation of average, but nothing else.

Just curious but is there an addon you can use to give you a timer on your internal cds? I roughly count it in my head on fights so I know when to use cds for more benefits, but it would be nice to have an actual number.

One example is Procodile but you will need some test dummy time for it to learn the procs.

Proculas is a great alternate to Procodile, with the added benefit of being able to simply tell the addon the ICD.

Greetings,

Longtime WoW player but new to Elitist Jerks, so I hope my post is where it should be and hasn't already been addressed in the multitude of pages in this thread!

I am a level 80 FB/TTW spec raiding mage who's in a guild that runs Ulduar 25 on a regular basis. I am slowly but surely trying to complete that "BIS" set and get the best possible item I can to maximize my dps.

Currently I am in the process of focusing on getting new boots and a new belt, and fortunately I am also a level 450 tailor - so there are two patterns of interest to me that I can craft for myself: Spellslinger's Slippers and the Sash of Ancient Power. Both of these items seem to be rated very highly (per the original thread post) and would likely be a great fit for my toon, but I can't afford to make both right now and before I commit to using up all of those expensive mats I want to make sure I make the best decision. Currently both my boots and belt are level 213 so each would be upgraded similarly.

My first question is that between the Sash of Ancient Power and the Cord of the White Dawn, I am a little confused as to why the former seems to regularly get the higher marks (assuming characters are hit-capped). It is my understanding that as a FB mage, the two most important stats are (generally) SP and crit. The "Cord" has both 48 more crit than the "Sash," as well as having the extra bonus of a spirit stat - which gets converted into even more crit thanks to Molten Armor (and the applicable glyph). Additionally, the Cord has yellow and red sockets - as opposed to the Sash's red and blue - which generally tend to be better for dps casters (one can get a SP + SP/crit gem combo, for instance). With all of these features, I am missing the reason it's not the obvious choice for a mage in this category over the sash.

With this in mind, I am considering crafting the slippers for myself instead (the next "upgrade" would be a drop from 10-man Algalon which we have not yet completed), which seems to make more sense as there is no other easily obtainable direct comparison for me. I could then use Emblems of Conquest to purchase the Sash of Potent Incantations from a badge vendor instead, which, while not as highly rated as the other two belts, is still a 226-level item and yields a whopping 100sp bonus (as well as spirit and crit, which again, the Sash lacks).

My question is really twofold here: why is the Sash rated higher than the other two belts given the importance that crit has for FB mages, and assuming I go ahead and make myself the slippers instead of the Sash, would the badges be well spent on the vendor belt instead of wating to save up for a tier piece instead? (Currently I have two pieces of 8.5 gear, but neither are the helm or the chest piece; the two items purchasable from a vendor. Right now my chest is level 213 while my helm is 219). I would include a link to my character's page on the armory, but it is unfortunately down for maintenance at the moment.

Thanks for taking the time to read my post and address these questions.

Could you let me know the Fireball spec BIS gear without Algalon? Thanks in advance.

If you can do all hard-modes you can do algalon (it may take a couple weeks to get people accustomed to the fight but you have everything you need gear and skill-wise) so this is a useless list.

hm.. that calculation isnt really good.. you have to calc (1x39sp instead of 23sp and 2x 39sp instead of 2x9spell/8spirit)
cause you need 2 blue gems for your meta :] so you have +76spell :]

The Eyes aren't prismatic anymore , they count as red gem. So you will need 2 blue gems anyway.

Enthorn, would you consider uploading your Rawr character configs in the OP?

I'm adding support for importing the Rawr configs into SimulationCraft to make it easier for me to ensure my modeling stays current.

Theres two things that completely alluded you there;
Number one is that BiS is termed to mean Best in Slot, but here on EJ it should be BIS, Best Item Set. That means each item is valued against the other gear you're wearing, not just the other items in that slot.

Number two is that when you take Best in Slot gear, almost all of the time, you are dropping the item set bonus's from being considered, meaning that set bonus's account for more dps gains (most of the time) than the best in slot gear does.

If you want to use just Best in Slot gear, and not care about how the rest of your gear works, MaxDPS.com - World of Warcraft Gear Rankings - News and WoW Heroes - World of Warcraft PvE character info & ratings are out there to stroke your e-peen with. Oh and MaxDPS never once provided useful results. It's the dummy check for idiots who are new to raiding that can't tell the difference between an Ass and a Donkey.

If you have been following this thread at all, then you would have noticed that from the release of Ulduar until now, there have been many changes as to what the optimal gear is for both FFB and FB. While you may be skipping over an item from early in the instance to wait for one from later in the instance, in the end, both are negligible because they are both not the most optimal for BIS. And are you forgetting about [Starshard Edge]? Since you talk about hard modes, and it's awarded from all Keeper/Watcher hard modes now too, why not use the same logic, and wait for that dagger to drop?

And what do you mean only one dps trinket? Last I checked, if you needed hit, it was a dps boost.

Correct me if I'm wrong on this but I believe that each of the "Algalon" weapons only drop from 1 other Hard Mode boss (in addition to Algalon) and not all of them. e.g. [Constellus] only drops from Hodir's Rare Cache, [Starshard Edge] from Firefighter, etc.

This is correct.

This is not correct.

This site is so hilariously useless it's completely puzzling how anyone ever referred to it. I guess it falls in the same category of retardation as Be Imba! - the online Character Auditor for World of Warcraft.

You're right, I should have reworded it better, but such things are easily skipped.

You know what I mean when I'm saying that. I don't mean that a best in slot item is ignored all together, it's just more optimal to find out which set of gear is the best, it's always been kind of obvious as to which items are best in slot vs best in set, given that you are capable of gearing correctly for hit cap. In the first post it refers to a full set of items to gain maximum dps, not personal gear vs. options available, that's what RAWR does.

You should try to avoid getting hit from your rings and trinkets if you plan on getting close to bis because then you don't rely on them as a crutch for the hit.

What? I think most players would agree that hit on rings/trinkets is not a crutch, regardless of whether your character is going for BiS gear or just trying to get whatever upgrades may be available to them.