Add Dark Thoughts tracking in some form.

Looks like there is a 10-stack (each dot per MF tick counts for 1) floor before a proc can occur, then it is 3% chance per DoT on the target per Mind Flay tick (according to tooltip).

Ideas to add:

• Floor threshold reached TRUE/FALSE
• Probability of a proc occurring during current channel or tick
• Confidence interval (80%?) that a proc will occur in the next X ticks / Y seconds / Z GCDs
• Others?
• Time remaining on the buff (why not)

Outstanding questions:

• Are proc rolls independent per DoT or additive for a single roll per tick?
• Any ICD? Bad luck protection? Does MF count as a DoT once they system says a proc can occur (via SWP/VT/DP)?

May end up just closing this and not supporting any prediction.

Reopening. Will implement tracking following the methodology SimC uses to grant procs:

x = 3% per DoT on target when struck, up to 3
x% chance per tick of Mind Flay
0.5 * x% chance per tick per hit of Mind Sear

This might follow something other than a binomial distribution but I'm going to model it as if it were until there is conclusive evidence otherwise.

Going to log values for a while and see where it ends up.

Add a configuration for the threshold to alert that a proc will occur "soon", based on a geometric distribution.

More research is needed to determine how these procs work from a probabilistic POV. It isn't exactly a binomial distribution and may be more like a beta-binomial distribution.

Not seeing any value behind adding this. Closing it off.

Early findings from my logs from a raid last night. This assumes Mind Sear is worth 1/2 a "tick" per DoT, or 1.5% chance per dot that is ticking on the target when it is struck. When a proc occurs, the counter resets to 0. All "dot tick proc" chances are counted against the total which may be a slight (~1-2) overestimation in some cases (see below about 9).

Minimum number of "dot tick proc chances": 9
Maximum number of "dot tick proc chances": 111
Sample size: 96
Average (Mean): 38.0365
Std Dev: 20.4709

I have a feeling, based on the 9s (3 of them) and a 10 (1 of them) that the way Blizzard tracks this is with each DoT getting it's own 3% chance roll to proc it after a minimum of 10 proc chances. None of the 9s or 10 that proc'd last night had any contribution from Mind Sear, only Mind Flay.

In this case to get a 9:

• Target has multiple DoTs on it (let's say 2, SWP and VT).
• Mind Flay tick occurs.
• SWP check happens, causing internal counter to go up by 1.
• Counter is >= 10, SWP roll happens and procs Dark Thoughts.
• Internal counter resets to 0.
• VT check happens, causing internal counter to go up by 1.
• Counter is 1 < 10, no VT roll happens

Now we enter in to a situation where we have 1 on the internal counter.

• Target now has all 3 DoTs on it.
• Mind Flay tick 1 occurs.
• DP check happens, causing internal counter to go up by 1 (2).
• Counter is 2 < 10, no DP roll happens
• SWP check happens, causing internal counter to go up by 1 (3).
• Counter is 3 < 10, no SWProll happens
• VT check happens, causing internal counter to go up by 1 (4).
• Counter is 4 < 10, no VT roll happens
• Mind Flay tick 2 occurs.
• DP check happens, causing internal counter to go up by 1 (5).
• Counter is 5 < 10, no DP roll happens
• SWP check happens, causing internal counter to go up by 1 (6).
• Counter is 6 < 10, no SWP roll happens
• VT check happens, causing internal counter to go up by 1 (7).
• Counter is 7 < 10, no VT roll happens
• Mind Flay tick 3 occurs.
• DP check happens, causing internal counter to go up by 1 (8).
• Counter is 8 < 10, no DP roll happens
• SWP check happens, causing internal counter to go up by 1 (9).
• Counter is 9 < 10, no SWP roll happens
• VT check happens, causing internal counter to go up by 1 (10).
• Counter is 10 >= 10, VT roll happens and procs Dark Thoughts
• Internal counter resets to 0.

If this is indeed how it is happening, then it should theoretically be possible to get an 8.

Going to continue gathering data.

Nightly update on this:

Sample size: 184
Average (Mean): 36.5761
Median: 33
Std Dev: 18.8478

No 8s. Some of the super high outliers could be bugged or due to zoning/reloads/relogs.

Interestingly, after 23 chances @ 3% per chance, we'd expect just over 50%, 1-((1-0.03)^23) = ~50.3694%
Since we can't get a proc for the first 10 attempts, our Median for attempts is: 33 - 10 = 23.
Coincidence?

Looking at the data where the proc and attempts to proc were from Mind Flay ticks ONLY (removing Mind Sear as a variable), there's some movement downward (could Mind Sear really be 1/3 of a tick instead of 1/2?)

Sample size: 161
Average (Mean): 35.09317
Median: 31
Std Dev: 18.06405