Originally Posted by jmichaelp
My playing observation, in-game experimentation & comparison in raids to other hunters indicated (for my situation at least, over a few months time) that although the spreadsheet indicates hit-capping will result in higher DPS, actual play showed differently. As alluded to in previous discussions, I think this anomaly exists due to differences between consistent DPS (from hit-cap) & the potential to have higher PS when not capped.
This conclusion was reached prior to the last two patches - have any changes occurred in these patches that might impact my previous conclusion?
Edit: After reviewing the spreadsheet, I don't think that hit is modeled as a smooth function. I also noticed that there's a "hit adjustment" in the calculations page, from which I conclude that hit couldn't be modeled just right otherwise an
adjustment would be unnecessary.
|
Your conclusion is unlikely, given the degree of applicability of hit to the full range of damaging attacks -- as long as you are considering average dps over a significant period of time.
Consider the mathematical model is based on the (vastly oversimplified) premise that:
99% of 100 damage == 100% of 99 damage
Trade-offs between hit and other stats come down to changes in the percent or changes in the damage. For example:
99% of 101 damage > 100% of 99 damage
This was the topic of earlier posts in this thread, which some folks found very difficult to swallow.
You seem to be asserting that (basically):
99% of 100 damage > 100% of 99 damage
Since the 99% is probabilistic, there will exist segments of a fight where you have no misses. For those segments, the equation looks like:
100% of 100 damage > 100% of 99 damage
Which is a duh, of course. But the random number generator doesn't "reset" at the beginning of each fight. Your first shot is just as likely to miss as your 100th. So you
will eventually have misses. Which will significantly pull down your average dps in the vicinity of that miss.
Now, given how proc effects and cooldowns interact, accurately calculating the full impact of hit is somewhat difficult. This was one of the contentions made by the "hit cap is #1!" crowd. They claimed that the problem was complex enough that the spreadsheet couldn't get it right, and therefore must actually be under-valuing hit. You would seem to be on the opposite side of that fence, with only anecdotal evidence on your side. You either need to demonstrate
mathematically why the model is wrong, or provide statistically significant (i.e., not anecdotal) data that contradicts the model for your claim to be well received.
Most likely, something else is at work:
a. You are drawing erroneous conclusions from the spreadsheet because you aren't quite using it correctly
b. Your sample size is insufficient or you are inadequately analyzing it
c. You are alliance and are forgetting about the 1% hit you get for free in raids (and thus are 1% over the cap most of the time).