Quote:
Originally Posted by Jimjam
[You must be logged in to view images. Log in or Register.]
I don’t think FSI is about mitigation anyway - opposite to you I’m a little surprised the calculated average benefit is so high. My perception of FSI value is the benefit it provides at the extreme end of the normal - where that bash stun initiates a sequence of catastrophic failure. Obviously, by definition, that is more of a fringe occurrence, but for someone who is failure adverse (perhaps they like to kill rare, highly competed mobs - something OP wonders whether lacking JBB may make a difference) that may be a decisive benefit. For me, someone who likes to deal with the spice of things going a bit pete tong from time to time it’s not very enticing.
|
Oh, I agree absolutely. FSI shows its value when shit hits the fan. "Initiates a sequence of catastrophic failure" - this is exactly why I bring up Markov Chains. With that sort of analysis you can encode the state of the fight, including what spells have landed and how much damage you have taken. For example, the state at the start of the fight will be full hp; no debuffs. The two possible next states will be Malo landed; Malo failed to land. You give each state transition a probability, perhaps 85% and 15%. This lets you build out the full state space of a fight. Some parts of the chain can contain loops, like Casting Slow -> Slow failed to land -> Casting Slow.
The value of FSI is that it changes the probabilities of those state transitions, making it less likely you end up in the state of low health, slow failed to land. Markov Chain analysis was invented for precisely these types of problems, and the wikipedia article is at least somewhat intelligible:
https://en.wikipedia.org/wiki/Markov_chain#Transitions
I'm not willing to put in that much effort for this topic, and so went with a simplified analysis. But you correctly pointed out the "wiggle room" I was talking about; DSM I think intuitively understand the objection as well, but unfortunately he is unable to articulate it, which is why he's now talking about wizard nukes. I think the next area of improvement would be to replace the DPS number with a distribution, which lets us recognize that sometimes over a short interval a mob will do drastically more than average damage and sometimes drastically less. If, for example, we can say that 5% of the time a mob hits you for 400 damage, we can further refine the range out outcomes, and better quantify the probability that a fight turns so poor that one is forced to abandon the effort and gate out.
Unfortunately, as a filthy casual who's never leveled past 51, and never played a tank at all, I simply don't have the logs or intuitive understanding to build that probability distribution, and so I must rely on the DPS number that was provided to me. But if anyone else is interested in that analysis, I remain willing, even eager, to collaborate.
Quote:
Originally Posted by Troxx
[You must be logged in to view images. Log in or Register.]
Do you work in higher education (undergrad, masters or doctoral level)? You carry yourself like a seasoned professor that is perhaps overly patient.
|
I greatly appreciate the kind words. My pedagogy has been developed through three activities - coaching novices in powerlifting, teaching the basics of tree identification on nature walks, and at work, where as a principal engineer, I spend a great deal of time mentoring junior engineers, some of whom have been quite dense.