Memory Blur Calculation
 

From my experience playing an enchanter on P99, I feel our Memory Blur calculation is incorrect. It favors a success roll too much here I believe.

Memory Blur success chance should be a sum of three values:

1. Level of Mob Factor
2. Spell Raw Value
3. Charisma Factor

Level Based Factor:

The way Devs described the first factor, is like this:

Mobs below level 17, have a 100% value to memory blur
Mobs above level 53, have a 25% value to memory blur

They are less clear on what happens to mobs inbetween that range, but I believe we can safely assume a linear model, in which case it can be represented by this function:

Level Based Chance = -2.08333 * Level + 135.4167

Spell Raw Value:

This is simply the value shown in the spell data for the actual memory blur component. For example, Blanket of Confusion has a built in 20% Memory Blur Raw Value (however it should be noted Blanket has 2 of these, meaning it should perform the memory blur roll twice, making it the most reliable memory blur available to Enchanters)

Charisma Factor:

This is simply (CHA - 150) / 10. Meaning a max (255) CHA enchanter would have a 10.5% value.

----------------

Example 1: let's use any ToV trash mob (excluding baby drakes), with an enchanter using Memory Flux, with 255 CHA.

1. Level Component for level 53+ mob = 25%
2. Raw Value for Memory Flux = 30%
3. Charisma Calculation for 255 CHA = (255-150)/10 = 10.5%

Sum = 25+30+10.5 = 65.5% Chance your Memory Flux spell should roll a success and blur the target.

Example 2: let's use any ToV trash mob (excluding baby drakes), with an enchanter using Blanket of Confusion, again with 255 CHA.

1. Level Component for level 53+ mob = 25%
2. Raw Value for Blanket of Confusion = 20%
3. Charisma Calculation for 255 CHA = (255-150)/10 = 10.5%

Sum = 25+20+10.5 = 55.5% Chance your Blanket of Confusion spell should roll a success.

HOWEVER! This check is performed twice. So the chance that either both rolls are a success or atleast one roll is a success is most easily calculated by:

1-(1-.555)^2 = 80.2% Chance that you Successfully memblur (basically check the chance of failing both rolls and see the difference from 100%)

Example 3: let's use a max level Illis Froglok in Sebilis, when using level 4 Mesmerize, and again 255 CHA.

1. Level Component for level 53+ mob = 25%
2. Raw Value for Memory Flux = 1%
3. Charisma Calculation for 255 CHA = (255-150)/10 = 10.5%

Sum = 25+1+10.5 = 36.5% Chance your Level 4 (or any mez due to them all having a 1% memblur) should roll a success and blur a target.

Evidence:

Naturally for something this specific you need developer quotes, and of course those aren't in era but something this inherent in the game code is unlikely to have been modified hugely over the games life:

http://www.eqclerics.org/forums/showthread.php?t=23401

Discuss.

It would be very easy to change memblur mechanics assuming that Verant wrote their code in a vaguely correct way.

Also, your numbers (80% for blanket and 35% for L4 mez) look pretty close to what I've seen on Loraen.

the modifier is gigantic, i'm really skeptical that was classic, particularly because the 1% chance on mezzes was very very uncommon on live from what i remember, as a mained enchanter back then. not that that proves anything of course, but similar to charm, it's one of those things where we've kinda accepted that no proof can be produced that would enact change.

i'm confused what point you are trying to make though dald - like raev said, I think in-game values reflect your stated expectations

Do you have any evidence that suggests the success rate is higher than your proposal? FWIW, if I had to guess, I'd say success seems less likely on p99 than the rates you've provided. 36% mem blur chance on at 53+ mob with lvl 4 mez, in particular, seems higher than my experience. With 36% chance, the expected number of casts for a success is less than 3. If that's accurate, I will be forgoing mem blur spells in favor of a few casts of low mana, fast cast mezes, save for when aoe mem blur is required.

Assuming the mem blur spelldat data hasn't changed between classic and the time of this evidence, it seems reasonable.

My personal experience is that the Mez line should be roughly 35% for mobs over 53, and I feel like it's way closer to 50%.

What you described Fash is exactly what I've gotten used to doing. Level 4 Mez is superior to most memory blurs assuming the mob is mezzable (I.E. 55 or lower). Which can be correct from a mana standpoint but from a success rate I don't think it should be firing as frequently as it does.

Also as a cleric, using Atone and not having CHA exceed 150, it should be about 26%, yet that seems to function more towards the 40ish% range.

Maybe it is working, I just thought it was worth a post to see if those with access to the code could take a peak and see if it does line up!

do you have any insight about Reoccuring Amnesia? what's wrong, how to fix it, etc.

I must admit, I found this post very interesting. Learn something new every day in this game.

Presumably, it should cause a memblur(25%) every tick for its duration (4 ticks). Currently it has no duration, so the memblur occurs only once, when the spell is cast (just like every other memblur spell).

I take that back. Testing on a lvl 49-51 mob with 241 CHA, success rate of lvl 4 mez was 44% (out of 96 tests), which is close to the expected, 41%.

I'm going to necro post this.

The cleric reference says the charisma stat is maximum 15 percent based on charisma this would mean that 300 charisma is effective even if Max visible charisma is 255 is this correct?

Has anyone tested this on a cleric? This would mean charisma basically has no impact due to so few clerics are going to go much beyond 150 without serious effort.

In the example calculation it says charisma - 150 / 10
For a 150 charisma player that would be 0 / 10 so zero.

However in the example it reads as if it was 150 / 10 = 15