AC softcap and low level hardcap discussion

[bcbrown]

Yeah, I'm doing a Monte Carlo simulation as well, written in Python. I was wondering if you had found a closed-form solution instead of simulating.

I get access denied when clicking the link to your source code, but I could view the source from the website you made. One difference is that it looks like you aren't casting to integer for some of the intermediate values (including the Roll0 return value).

The other difference is that I was calculating the average value of the index while you're calculating the average value of the modifier. Once I started adding 1 to the index to account for that I get these values:

Mitigation: 100, Wrath: 100, average roll: 10.43
Mitigation: 200, Wrath: 100, average roll: 6.21
Mitigation: 400, Wrath: 100, average roll: 3.98
Mitigation: 800, Wrath: 100, average roll: 2.99

Pretty close. If you're curious about the code, it's this:

 

from random import randint

def roll(max_value):
    if max_value - 1 > 0:
        return randint(0, max_value - 1)
    return 0

def approximate(offense, mitigation):
    atk_roll = roll(offense + 5)
    def_roll = roll(mitigation + 5)
    avg = max(1, int((offense + mitigation + 10) / 2)) 
    index = max(0, int(atk_roll - def_roll + avg/2))
    index = int(20*index/avg)
    if index < 0:
        index = 0 
    if index > 19: 
        index = 19
    return index + 1 


def monte_carlo(offense, mitigation, num_rolls = 100000):
    results = 0 
    for i in range(num_rolls):
        result = approximate(offense, mitigation)
        results += result
    avg = results/num_rolls
    return avg 


if __name__ == '__main__':
    offense = 100 
    for mitigation in [100, 200, 400, 800]:
        avg = monte_carlo(offense, mitigation)
        print("Mitigation: {}, Wrath: {}, average roll: {}".format(mitigation, offense, avg))

Originally Posted by DeathsSilkyMist [You must be logged in to view images. Log in or Register.]

I've been thinking about this. I tried a few ways to write some succinct math, but honestly I think most people glaze over the formulas when they have to calculate it themselves.

Yeah, almost no one gives a shit about the math, but I don't blame them either. It's almost completely irrelevant to actually playing the game. Most people can just follow these three guidelines and do just fine:
* More AC, more gooder
* HP > AC for raid tanking
* If you want to know which weapon is better, parse both options against a mob comparable to the mobs you'll be fighting against. Or ask someone who's already parsed them.

I find understanding the math and the internal mechanics interesting in its own right, and I understand it a lot better than I did a year or two ago, but I wouldn't say it's made me better at the game or at deciding on itemization or anything. Just a fun hobby.

[Goregasmic]

Originally Posted by bcbrown [You must be logged in to view images. Log in or Register.]

I find understanding the math and the internal mechanics interesting in its own right, and I understand it a lot better than I did a year or two ago, but I wouldn't say it's made me better at the game or at deciding on itemization or anything. Just a fun hobby.

Yeah. If the EQemu softcaps are right, outside of TOV loot i feel like chain/leather doesn't have a ton of options to optimize anyway. For rangers it is hard enough to get decently balanced stats without dipping below 200ac. Rogues probably have a bit more leeway as dex doesn't seem as important to them. For MTs it is already solved with going pure HP anyway and for cloth I'm not sure any amount of AC matters terribly. It probably gets more interesting for knights and bards as they have more stats to deal with where potential excess AC could go to. Could possibly matter a bit for druids who have a more hands on play style if the cap is 200 but at 120ac I guess you reach that without trying? Probably worth keeping an eye on for monks too.

I guess the main problem is itemization. Most gear slots have a given BIS that is head and shoulders above everything else and it is kinda hard to sacrifice those for a given stat. When you reach the TOV gear stats feast you have a bit more options.

If those caps are valid during vanilla/kunark it could be a lot more meaningful for any class that gets punched in the face due to the gear not being nearly as stats inflated as velious options, if we ever see that era again.

[DeathsSilkyMist]

Originally Posted by Goregasmic [You must be logged in to view images. Log in or Register.]

I guess the main problem is itemization. Most gear slots have a given BIS that is head and shoulders above everything else and it is kinda hard to sacrifice those for a given stat. When you reach the TOV gear stats feast you have a bit more options.

If those caps are valid during vanilla/kunark it could be a lot more meaningful for any class that gets punched in the face due to the gear not being nearly as stats inflated as velious options, if we ever see that era again.

Originally Posted by bcbrown [You must be logged in to view images. Log in or Register.]

It's almost completely irrelevant to actually playing the game. Most people can just follow these three guidelines and do just fine:
* More AC, more gooder
* HP > AC for raid tanking

I would say there has certainly been some nuance discovered with AC testing that does directly affect how players play. It may not matter as much on our lower population, top heavy servers. But there are still players who have modest gear, and are looking for their next upgrade.

1. The low level AC Hardcap shows that you shouldn't focus as much on AC at lower levels. If someone twinks a level 1 Monk with the same gear as my Monk's Magelo in my signature, they would have ~140 worn AC. This means they wouldn't have to worry about the Monk Weight Limit until at least level 21. This is because the level 20 AC Hardcap is (20 * 6) + 25 = 145 worn AC. The Monk Bonus AC is being Hardcapped from levels 1-20 in this scenario, so there is no reason to stay under the Monk Weight Limit. Carry as much stuff as you want.

2. The confirmation that shields increase your AC Softcap make them more important in some scenarios. At low levels shields are the only way you can increase your AC a little bit once you are Hardcapped, as an AC Softcap still exists at some point before the AC Hardcap. This also makes items like Yakatizma's Shield of Crafting more desireable. Before I was hesitant to get one, because I still use the resistance shields like Shield of the Red Dragon to cap specific resists. I didn't want to spend a ton of DKP on an item in a slot I am still swapping out. But Yakatizma's shield increases your AC Softcap by 75. This means Shamans wearing Yakatizma's shield are probably able to mitigate the damage taken from hits about as well as a semi-raid geared Shadowknight/Paladin.

3. There is some nuance to -AC items like Hammered Golden Loop. Before I did not like this item as much, because wearing two would reduce your worn AC by ~40 when you include the positive worn AC you get from wearing other earrings. That is basically a breastplate's worth of worn AC. But if you are well above the AC Softcap for your class, the -AC doesn't matter that much. For example, it looks like a Shaman gets ~17% returns on worn AC past the AC Softcap. Losing ~40 worn AC would really only be a loss of 40 * 0.17 = ~7 worn AC if you are still above the AC Softcap while you have them equipped.

4. When you know your class's AC Softcap, it is easier to know what worn AC value you should aim for. AC Softcap returns aren't great on most classes, so it is probably better to focus on other stats once you hit your AC Softcap. That doesn't mean AC is worthless past the AC Softcap, it just means you need a lot more AC past the Softcap to see improvements. You are basically looking for Velious raid gear at this point, which has high AC and great stats.

Originally Posted by bcbrown [You must be logged in to view images. Log in or Register.]

* If you want to know which weapon is better, parse both options against a mob comparable to the mobs you'll be fighting against. Or ask someone who's already parsed them.

My DPS calculator in my signature is great for this, as you can quickly get a rough estimate for which weapons are better in a short period of time. It can be difficult to get reliable data from players in my experience, especially if the weapon in question isn't widely used. Getting good parses yourself can take a decent amount of time, so parsing like 10 different weapons in game will be time consuming. This also assumes you have the weapons to parse them in the first place.

[Snaggles]

Originally Posted by Goregasmic [You must be logged in to view images. Log in or Register.]

Yeah. If the EQemu softcaps are right, outside of TOV loot i feel like chain/leather doesn't have a ton of options to optimize anyway. For rangers it is hard enough to get decently balanced stats without dipping below 200ac. Rogues probably have a bit more leeway as dex doesn't seem as important to them. For MTs it is already solved with going pure HP anyway and for cloth I'm not sure any amount of AC matters terribly. It probably gets more interesting for knights and bards as they have more stats to deal with where potential excess AC could go to. Could possibly matter a bit for druids who have a more hands on play style if the cap is 200 but at 120ac I guess you reach that without trying? Probably worth keeping an eye on for monks too.

I guess the main problem is itemization. Most gear slots have a given BIS that is head and shoulders above everything else and it is kinda hard to sacrifice those for a given stat. When you reach the TOV gear stats feast you have a bit more options.

If those caps are valid during vanilla/kunark it could be a lot more meaningful for any class that gets punched in the face due to the gear not being nearly as stats inflated as velious options, if we ever see that era again.

Outside grind groups and solo challenges AC does almost nothing for all but a few classes (who are subject of a CH chain). Even then, for hybrids having a deep mana pool for those stunt kills along with decent dps/damage shields is more important.

I feel SV’s are under prioritized in general. Outside a few targets that you won’t resist the AE (and it’s a doozy where it becomes a wort click battle), cranking SV’s to the moon helps over HP’s. Get the cheap flowers too; CR and the flower stack as well as PR and the other. The disease flower is basically free.

I know this is a segue to an AC topic and may start a new thread to discuss. I just feel people aren’t tracking svs and I have anecdotally seen the benefit. I also had shamans shame people while praising how few torps I have needed.

[Goregasmic]

Yeah for raiding I guess most classes get pulverized if they get aggro from a raid target unexpectedly with no way to mitigate it (fd/disc/idol/whatever).

I enjoy raiding every now and then but it always seems to quake at stupid hours for me. I'm more of a fan of low man stuff where gear absolutely makes a huge difference in most cases so I kinda like tuning for that. In that scenario decent MR is basically all you need but getting softcapped on AC is kinda helpful and also stuff like dex for procs who could make up for suboptimal group comp. Outside raiding SVs are kinda MR > frost/poison > disease > fire... and MR is ahead by a long shot.

But from what I read it seems like resists are broken on p99 (spells land much more than they should) and it seems to track. At 225MR i get feared all the time and in HOT on red wurms at 200+ fire i still get hit a lot. One night I forgot to put my FR on and I did notice a difference but it was not very significant. Doesn't make we want to go that extra mile and pay like 40k for flowers and I know a lot of people feel that way but it isn't like I've parsed that stuff either. Also, when you're not a raider swapping to resists usually means replacing solid pieces of gear for resist junk so the trade off kinda hurts even more.