Modeling Enmity

[kanican]

Overview

I don't claim to know how everything in this game works.  There are so many complex equations dealing with just about every part of how this game functions that it's really overwhelming to just think about it.  There are much better, more "famous" players to look to for equations like Aurik or VZX.  I do, however, believe that I can tell

Comments

"beauty"

[anonymous]

If finding an explanation like that beautiful makes you a geek you're in good company.

"It must be beautiful"

Nice work, anything with functions and graphs gets my thumbs up :D

-random LJ fan.

[anonymous]

i work by a slightly different theory for a few reasons

[kanican]

I think as someone mentioned earlier, you don't really need calculus to do this, I just happened to write it in that form (was pretty much the simpliest type you could do too

[anonymous]

It would be nice if you could do more testing from the PLD hate standpoint.

Or the Redmage for that matter.

[kanican]

My observations come mainly from NIN/DRK tanking. My LS rarely uses RDM/NIN or PLD/NIN so I feel I am not quite suitable to talk at length about those tanking styles. I've certainly seen them and have had them used in LS, but not nearly to the extend of NIN/DRK. That's why my examples are geared towards NIN/DRK.

!!

[kainsin]

Dude, gratz on the Novio!!

Re: !!

[kanican]

Thanks Kain!

[amlaruil]

Reducing hate at an asymptotic rate doesn't require calculus. A simple equation would be:

every time period do: hate = .9*hate

One thing you don't address is the possibility that actions a mob does to you reduce hate. A specific example of this is Ghrah; if you aggro one (but don't cast on it) it'll lose aggro after 2 actions that fail to hit you. Of course this doesn't seem to match any of the normal observations about enmity, so it may be a different mechanism altogether. Other evidence that taking damage reduces hate is difficult to differentiate from enmity decaying over time.

[amlaruil]

you could do 10% hate lost/unit time but you would still be left with the problems of hate reaching larger numbers than you want to store in memory if the player committed enough actions in a short time - an example is a weak mob like oni carcass which has enough HP that you could relic barrage > etc. and the unit time for hate loss would have to be under 1s to avoid using more than 1 byte~ and you don't want to be calculating hate loss constantly because that's heavy on processing.
another problem with a decay function like that is how hard it is to maintain higher levels of hate (which nin/drk generally shows can be maintained), if this was the case a blm nuking would lose hate almost immediately unless the unit time was very long (ie. around the 5-10s mark) which we can observe is unlikely.
overall it would be difficult to get the correct function (0.954, 0.831, 0.9206....) compared to giving static hate loss/unit time.

it could be done that way of course, i personally dont think it is.

[kanican]

I was looking more for a qualitative model for this. I think it would be pretty hard to get anything concrete for a quantitative analysis on this since we can really only measure hate in a relative term, not in absolutes.

To me, the importance is to develop a model that can somewhat accurately predict what would happen (for strategy), not necessarily figure out what SE really does to calculate hate. Even if said model does not work, if it accurately describes the vast majority of cases, I think it still has tremendous use.