I don't have the notes on the math anymore, this was pretty much a cut and paste of the conclusion post that I put on our boards after I finished the analysis. Keep in mind that while the "high DI" (15-20 DI) probability may be X%, the high round chance is not X^4%, because you could have a 20, 20, 10, 10, for example, and thats a "high DI round" according the the criteria of the mob DI/DB distribution. I spent quite a while figuring that out, but if you want to repeat it yourself or check new values or parse values versus hypothetical ones, you'd need the distribution on DI and then calculate the probability (note again that I was assuming no avoidance, it was bad enough trying to do this by hand 15 years after my last statistics or calculus class) of getting 60+ cumulative DI over 4 hits. It was also based off a hypothetical average DI on a pal/sk of 5.0 (which would be true for "some" mob out there, but not necessarily mobs in any particular zone or raid mob). It was just an example and starting point to illustrate the relative probability difference. Your rAC (ignoring buffs) versus another persons is going to give % returns on the rAC gap in terms of DI shift that is fairly linear with the rAC gap, as the entire point of calculating rAC is to remove the effect of the softcap creating the geometric curve on ac to DI shift returns. I really should have included all that stuff now that you mention it, but the original post was meant more to summarize with some background info for people in my guild, not for broad dissemination to parse hounds. If I ever feel like it I'll take the time to rework it and post the details, but I'm betting someone else could easily do it and probably has more free time than me.
AC and HAgi augs may help some, sure.. but the chance of any caster surviving a round or 2 from a boss mob is pretty low. Guess someone can parse that out if they want, but I'm not going to pick all AC or HAgi augs over a good balance of HP / AC / mana, HPs being more important for me so I can survive the AEs and such.
Did you measure hit distributions or calculate them by AC? Your first post implies a calculation based on knowing AC values following as it does a long exposition of AC numbers. Wouldn't a calculation like this require knowing the algorithms used inside EQ to randomize hits? How else could you get language like the following? A brief calculator job suggests to me your formula has average DI being inversely proportional to AC. AvgDI = mob offense factor / player effective AC That corresponds to the ratio of hits you give between the classes and their AC values. 6.5 / 5 = 2600 / 2000 Why do you believe this formula is correct? Do you have measurements to support it? You must know it is obviously wrong for small values of AC because DI can not exceed 20.
It's definitely not a linear relationship as I've charted it at DI1 and DI20. However they're accumulative as DI20 I guess means DI20+DI21+DI22+... etc
I don't believe Abazzagorath is offering up a formula per se, just very rough numbers to show how things work generally. In the past (I dunno how things are now) your AC did result in an average hit, or if you prefer DI value. If you plotted the distribution of hits you discover that they were a bell-like curve centered on that average. Lower your AC (a lot so that you can see the result) and the bell curve moved. At one time your AC could be high enough that you saw few if any DI=20 hits. It can definitely be low enough that you see all DI=20 hits. Sit (effectively setting your AC mitigation to 0) and you take all maximum hits. Stand with your armor off (low AC) and you'll still take very close to all maximum hits. There's not a whole lot of difference between sitting and say 2500AC vs. current content mobs. DI can't exceed 20, but it can be at 20, Damage vs. AC looks linear over a shortish AC span (a few hundred AC) but it's not linear.
Like I said in the same post that you quoted (but maybe I didn't make clear), I used a DI distribution chart from a log sample (using adds from fumerak) from some old logs. Mob attack versus your ac relative gap will always change based off expansion, mobs, you're gear, etc. That was just a representation of that one granular point. Fight a much harder hitting mob, your probability of a high DI hit is obviously higher, go fight orc pawns and its obviously lower. None of that really matters, because the point of the calculation was only to show that the probability of a high DI round was that much higher for the classes that have lower post-cap returns. If I am fighting something where my chance of a high DI hit is 50% of what it was before, the high DI round probability is reduced at a relative rate for someone that has less or more probability as me on the other mob. You don't need to know the algorithm the game uses. DI shifts scale linearly with rAC gains (every parse I've ever done in the past showed this, not that I've parsed it in a long time, but no reason to think its changed - assuming you do it right and ignore DI(1)). That's the entire reason they removed the AC hard cap and put in a softcap, instead of removing the cap entirely prior to PoP. I don't get why you'd think it works any differently for small changes in ac. If player A has 2450 rAc and player B has 2500 rAC, you'd expect player A to have an average DI that is 2% higher. DI not exceeding 20 is irrelevant. The only issue is with extremely LARGE amounts of AC because you can't shift DI(1) lower, and I ignored that because, again, it wasn't really the issue I was targetting when I went through that. If you really wanted to get into it, you'd exclusively focus on frequency of DI hits from 10-20 and how much shift occurs there (since you shouldn't care about low DI hits), and come up with three meaningful values 1) frequency of hits from 10-15 and 16-20 2) average DI of your highest 50% of hits. If you're looking for some formula to value AC versus heroics or Hp or whatever, this really wasn't the point. The point was to 1) give some background data and formulae, and 2) to make the point to some of our casters that using high ac augs with little hp/mana was a waste because their survivability was barely nudging, but that improvement was at the expense of their sustainable dps (mana regen from heroics or pure mana) or their survivability from AEs (hp). This was a repost, hoping it would be of help, and it was never some all inclusive study of the eq melee dps mitigation system.
Remove DI(1) and DI(20). Chart a hit distribution for two parses with divergent AC levels. You can't use average DI, the returns are linear in terms of DI(-1) shifts, meaning DI(1) (because you can't go to DI(0)) and DI(20) (because a big chunk of those are not affected enough to slip to DI(19)) are clouding your conclusions. My example was just that, an example. All the values (mob values, average DI, etc) were all hypothetical. AC returns are linear for rAC, %-wise (obviously, gaining 100 rAC from 1000 to 1100 is not the same as going from 2000 to 2100 for dps reduction), in terms of probability of a downshift.
As I recall ... The order of defensive skill checks is Block (for Monks and Beastlords) / Parry (for everyone else), Riposte, Dodge, Hit/Miss. The SoF AA's that allow for a special attack upon a Riposte (such as kicking or backstabbing or bashing) changed the order slightly, putting it to Riposte, Block/Parry, Dodge, Hit/Miss. AA's that allow for blocking with a shield or a weapon come between Dodge and the Hit/Miss check. So, a Wizard using the Staff Block (or whatever it is called) AA would have Dodge, Staff Block, Hit/Miss. A fully AA'd Monk using a 2h weapon would have Riposte, Block, Dodge, Staff Block, Hit/Miss.
There are two ranges here. One would be large shifts in DI where this simple model must obviously break because DI can't go over 20. I don't think this one is all that interesting. Obviously simple models can't go that far and mobs we face in game don't hit that hard. Lets ignore this range. If your model for DI vs. AC works in the range of 4 to 8 average DI players get hit for it is useful. No need to cover sitting or naked or 1st levels vs. Veil of Alaris raid mobs. The range I am interested in is player AC. Your calculation is a very simple inverse to AC. DI = mob offense / player AC. It means a newbie character who is getting hit for DI 5 and puts on 25% more AC takes only DI 4 on average the same as a Luclin character or a Veil character. This is much simpler than the sort of local linearity one expects in unknown but smooth functions. It says a great deal about the insides of EQ calculations. If I wanted to test that I would first need a good source of internal AC numbers, something you have clearly put a lot of work in. Then I would need parses testing a broad range. At each range of AC I would need an appropriately scaled test mob and at least two long parses establishing a base DI and the change with modified AC. It would be quite a bit of work as mitigation parses gain precision slowly. How did you measure it? What range? How precise was it? No, I am not looking for a master's thesis with color glossy photos, circles and arrows. Just some words of how you did it. It is a lot harder than just running some parses, or at least the sort of "ran some parses" one typically encounters on these boards.
Ok I get it....and it makes my head hurt to be honest. Silkies get in the back where you belong . But anyway im curious on Hstats. Defensive purposes for Hstr gives armor bonus while shield is equiped, Hdex increases chance to reposte, block and parry, Hagi increases chance to dodge and suposedly gives a "bonus" to defensive skill. For tanking classes which should we all focus on? If a tank also want to consider dps in this wouldnt Hstr be the way to go over Hdex and Hagi?
every cluster of 25 hdex is 0.8% riposte and parry from the parses on TSW for a total of 1.6% mitigation. every cluster of 25 hagi is 1% dodge from the same parses. Hstr bonus seems broken. Hdex is the best in terms of the greatest boost, however there are less hdex augs with high ac than hagi. It is easier to get clusters of 25 hgai while maintaining high ac. Shoot for the hagi augs and work hdex as your secondary, if you are planning on maintaining your aim at higher ac values.
