Calculating affliction rates (check my logic?)

Macian · December 2012

Howdy,

One of the ways I'm getting my head around class changes is figuring out old vs. new affliction rates, but it's late here, and I'm not sure if my process works. Would someone do me a favor, look it over, and let me know if I'm messing up?

Let's take Sentinels as an example.

Sentinels have an affliction speed (I'm told) of 2.35 seconds with Dhuriv.

Darts trap hits at 3.5 seconds.

To figure an affliction rate, we need to figure out how many seconds it takes them to hit with one affliction (to me, this is a cleaner number than the number of afflictions they hit with every second).

Step 1: Calculate the number of "hits"

We need to find a least common multiple for 2.35 and 3.5.

LCM of 2.35 and 3.5 is 164.5 (I made this easier by finding the LCM of 235 and 350, then dividing everything by 100).

2.35 and 3.5 should divide evenly into 164.5.

Over 164.5 seconds, you hit with a Dhuriv combo 70 times.

Over 164.5 seconds a darts trap will hit 47 times.

Step 2: Calculate the number of afflictions

Dhuriv delivers 2 afflictions - 70 x 2 = 140 afflictions in 164.5 seconds

Darts delivers 1 affliction = 47 afflictions in 164.5 seconds.

Dhuriv + darts = 187 afflictions in 164.5 seconds.

Step 3: Convert that into a number of seconds per affliction.

187 into 164.5 = 0.88

New Sentinel affliction rate while in a darts trap: One affliction every 0.88 seconds (rounded)

To help me compare, since last time I fought, they had quickfoot:

Sentinel with quickfoot = 1.89 seconds dhuriv combo = 2 into 1.89 seconds = 0.95 seconds (rounded)

E.g. sentinel in darts is marginally faster (8%) at afflicting than a Sentinel outside of darts used to be.

Am I messing up here?

(Sentinels, this is an example, not a complaint)

Xarian · December 2012

The way that I do this is a good deal simpler, but you're on the right track. I had a post on oldForums, but it probably won't be carried over so I'll flesh it out a little more here.

The two major numbers that we're concerned with are the "affliction rate" (how many afflictions per second a class deals) and the "cure rate" (how many afflictions per second they can heal). These are both fairly easy to calculate.

In general it's (number of afflictions done on a single balance)/(balance cost)+(number of afflictions done on secondary balance)/(secondary balance cost)+...etc

For a sentinel the affliction rate would be:

The cure rate is calculated similarly. Herb balance is 1.75 seconds on average, focus is 5 seconds, tree is 10 seconds, renew is 20. So the generic cure rate for focusable herb affs is:

The "rate" at which sentinels beat that cure balance is simply the difference of the two:

, or about 0.262969

Well, that's not a hugely helpful number. It tells us some things, like that the sentinel is beating curing (since it's positive, the affliction rate is greater than the cure rate). What would be more helpful is if we could tell how many seconds it took for a sentinel to "stick" an aff. Well, this 0.26 number is in terms of (afflictions)/(time). If we take the inverse, we'll have the time it takes to stick an affliction, in seconds since all of our time units above were in seconds.

is about 3.8. That means every 3.8 seconds, a sentinel will stick a new aff. However, since sentinels have access to impatience, as well as a stack above impatience, it's worth checking what the numbers are without focus. The handy thing about my formula is that it's very easy to edit and modify, you don't need to find the LCM or anything, just edit or remove the appropriate numbers.

Without focus, the formula becomes:

which is about 2.16 seconds.

This formula can even be modified to accommodate rebounding, simply by modifying the aff and balance times to account for more "rounds" of combat. For example, if you guesstimate that on average, a sentinel will get in 3 hits without rebounding and then need to use ambush on the 4th hit, the formula becomes:

or about 2.22 seconds.

I've been playing around a lot recently with how accurately this theorycrafting works out in practice, and I'm pleasantly surprised with how well it models actual fights.