Guide Spell Steal Efficiency Analysis

Something I just thought of: how would you take into account archer’s melee? Since all you have to do is hold down right click, I’d assume cps would be in the hundreds.

 

Archer melee attack rate is actually lower at fast+ speeds if you hold down instead of spam click.

 

For the purposes of this analysis, it doesn't change anything for the left side, up to around 6-8 cps (depending on where you want to stop casting more ice snakes and heals, or whatever). For the right side of the chart, past that point, it would increase efficiency in general (for probably about 70-90% of cases) but of course the spells per cycle curve would flatten out.

 

what is cancel stacking

 

wow... that's a lot of numbers
yep here is another math thread by druser to link

 

Ok, after soloing the Eye using a fast hybrid tank shaman, I gotta say that mana steal could really use a form of bad luck protection; there were multiple times where I was stuck spamming melee attacks on the eye for like ten to twenty second stretches without mana steal proccing at all, and it gets kind of obnoxious, especially when using items with -mana regen such as the crossing, black and salpinx

 

> fast hybrid
> shaman

(those are the main issues here)

 

That’s kind of the point.

The fact that fast hybrid is a problem is kind of a problem in and of itself; if the only way to get consistent mana recovery with mana steal is to drop your attack speed down as much as possible, that forces people to use items like Atlas and heavy melee items to get some modicum of consistency. My point is that that shouldn’t be the case. It restricts build diversity.

I’m just saying that you shouldn’t have to be forced to run a heavy melee hybrid just to get consistent mana steal.

And as for shaman, the existance of other issues in a build does not disprove the main point I’m making. I’ve had similar problems on assassin and archer before; shaman is just my most recent attempt and the one that hurt the most from it.

 

no, in theory, no matter attack speed, stealing will proc every 4 seconds of perfect melee
however, for hybrid, this massively reduces the time you can spend casting spells

problem is that wynncraft's spellcasting system is shit, and since we don't have one-button spells, we are forced to spend way too fucking long on that shit

 

In theory, yes, but since it’s based on percentages, and not a timer or hit counter, it’s possible to go for much longer without getting mana steal to proc. It’s essentially a smaller scale version of the gamblers fallacy.

What I’m suggesting here is literally something that’s just “used melees for four seconds without getting mana steal to proc? Your next melee is guaranteed to proc it, or has an increasing chance to proc it”. It would give some much needed consistency to fast hybrid mana steal but wouldn’t make it better in most normal cases

I feel like I should really say that I’m not talking about fast hybrid’s viability as a whole; I’m trying to address the edge cases of bad luck because those ended up being the largest problems in my experience

 

theoretically you can also have the same bad luck on super slow, but i guess a bad luck protection wouldn't hurt

 

You've got your issues mixed up. Faster speeds are actually more consistent than slower speeds, per the central limit theorem (more samples will decrease the variance). The biggest issue is probably shaman's split streams splitting proc chance, but another secondary issue is of course fast hybrid in general.

Fast hybrid is not fixable in the manner you describe, because of how the efficiency varies by different factors. Any such adjustment would end up strongly biasing for certain cps levels*.

*if you read my original fast hybrid stream-of-conciousness description with Giga from September or so, you can see how I've addressed this.

Of course, past all of these issues is the chance-based aspect of mana steal, which is what fundamentally causes variation in mana. Personally I would prefer it to stay this way, as it pushes on-the-fly adaptation as players push their dps higher, and also helps in the risk-reward balance.

 

Not necessarily; I described what a potential form of bad luck protection would be like:

Since melee speeds are capped, increasing cps wouldn’t actually affect the speed or amount that the bad luck protection triggers

The thing is, on-the-fly adaptation only works if the game gives you tools to adapt with. Randomly making players only able to use melee attacks doesn’t really do that, especially on the edge cases where bad luck hits hard. It feels more like you’re fighting against your own build than the enemies. With bad luck protection inconsistency would still be there (four seconds of just melee is far from optimal for most builds), but it wouldn’t define the entire build and playstyle when you have a string of really bad luck, which is what I take issue with.

That’s a big issue that needs addressing as well, but like I said, I’ve still had issues with mana steal on archer and assassin. It’s just much worse on Shaman because of this

 

Maybe get the correct stats or probabilities of proc-ing? To demonstrate heavy steal or fast stealing
Will the Poisson distribution help?

 

The proc distribution should just be a binomial.