I’ve always wondered, do think evade is calculated as a Dependent or an Independent variable?

BadPullsMarco

In case you are unfamiliar with the terms, in statistics an independent probability essentially means a probability we’re the actions of a prior event do not impact the probability of a future event, and example would be a coin flip.

A dependable probability on the other hand occurs when a prior event directly has an impact on a future one. An example would be the probability of picking a red skittles out of a bag. The more you eat the skittles, the higher the chance the next skittles will be red (there are less choices left that are not red).

Personally I think it’s a dependent probability working in a way similar to this,

If(not evaded){
Increase chance by x%}

I say this because it almost always evades during the last hit of a 5 combo, right before you autonomously through your special.

Could be in my head of course, it’s easy to fixate on the annoying moments until you see patterns that aren’t there but what do you guys think?

BadPullsMarco

Sorry about the typos, quick thought, typed on my phone.

captain_rogers

Interesting take
But I think it is an independent variable
The chances of getting a tail in 5 coins is higher than the chance in 2 coins, cuz of the increase in sample size.
You can apply the same logic here.

PantherusNZ

I say independant - why do you notice the evades on the fifth hit of a 5-hit combo? Because that's the evade that screws you up more than any other, so you notice it more.

Crcrcrc

I'd imagine the programming is more like this, with no actual calculation for how many hits have landed.

if playerhit=True:
(code for rolling for rng bc I'm too lazy to type it out)
if rngresult == (whatever the evade chance is):
defenderevade = True
else:
defenderevade = False

(Obviously this is way oversimplified but I'm not a game dev nor do I know what their actual code looks like)

As for the 5 hit combo thing, you only remember it the most because it's the only time that screws you over. More often than not they'll evade mid combo and you'll be able to parry their retaliation or dex it.

BadPullsMarco

Yeah for

Crcrcrc said:

I'd imagine the programming is more like this, with no actual calculation for how many hits have landed.

if playerhit=True:
(code for rolling for rng bc I'm too lazy to type it out)
if rngresult == (whatever the evade chance is):
defenderevade = True
else:
defenderevade = False

(Obviously this is way oversimplified but I'm not a game dev nor do I know what their actual code looks like)

As for the 5 hit combo thing, you only remember it the most because it's the only time that screws you over. More often than not they'll evade mid combo and you'll be able to parry their retaliation or dex it.

Yeah for sure, I think I mentioned that in my original post, but my gut says it happens disproportionately on that last hit. I wonder what Dr. Zola would say @DrZola

Crcrcrc

BadPullsMarco said:

Yeah for

Crcrcrc said:

I'd imagine the programming is more like this, with no actual calculation for how many hits have landed.

if playerhit=True:
(code for rolling for rng bc I'm too lazy to type it out)
if rngresult == (whatever the evade chance is):
defenderevade = True
else:
defenderevade = False

(Obviously this is way oversimplified but I'm not a game dev nor do I know what their actual code looks like)

As for the 5 hit combo thing, you only remember it the most because it's the only time that screws you over. More often than not they'll evade mid combo and you'll be able to parry their retaliation or dex it.

Yeah for sure, I think I mentioned that in my original post, but my gut says it happens disproportionately on that last hit. I wonder what Dr. Zola would say @ DrZola

Out of curiosity I did a practice match against nightcrawler using Thanos (so he didn't affect evade chance). I landed 30 total hits, of those 10 were evaded. Of those 10, 8 were from the first 4 hits of a combo and 2 were from combo enders. Funnily enough that's actually a perfectly even distribution with no bias.

Another thing I thought about was since most people start combos with parries, that's 2-3 hits that cannot be evaded at the start of a combo due to the stun. So for most basic attacks the chance of evading on the last 2 hits of a combo is much more likely.

BadPullsMarco

Crcrcrc said:

BadPullsMarco said:

Yeah for

Crcrcrc said:

I'd imagine the programming is more like this, with no actual calculation for how many hits have landed.

if playerhit=True:
(code for rolling for rng bc I'm too lazy to type it out)
if rngresult == (whatever the evade chance is):
defenderevade = True
else:
defenderevade = False

(Obviously this is way oversimplified but I'm not a game dev nor do I know what their actual code looks like)

As for the 5 hit combo thing, you only remember it the most because it's the only time that screws you over. More often than not they'll evade mid combo and you'll be able to parry their retaliation or dex it.

Yeah for sure, I think I mentioned that in my original post, but my gut says it happens disproportionately on that last hit. I wonder what Dr. Zola would say @ DrZola

Out of curiosity I did a practice match against nightcrawler using Thanos (so he didn't affect evade chance). I landed 30 total hits, of those 10 were evaded. Of those 10, 8 were from the first 4 hits of a combo and 2 were from combo enders. Funnily enough that's actually a perfectly even distribution with no bias.

Another thing I thought about was since most people start combos with parries, that's 2-3 hits that cannot be evaded at the start of a combo due to the stun. So for most basic attacks the chance of evading on the last 2 hits of a combo is much more likely.

That’s actually a fantastic observation with regards to the stun. That might be what drives the illusion (at least for myself) why combos seem to end disproportionately with an evade.

As for the Thanos fight, even though that’s a very small sample, 20% on the fifth hit is almost comically compelling…

klobberintyme

Simple exercise: use a true strike champ against any evade champ. Note instances of when Evade Failed pops up. My experience has been it really does seem random, especially with science ultron drones in aq.

BadPullsMarco

klobberintyme said:

Simple exercise: use a true strike champ against any evade champ. Note instances of when Evade Failed pops up. My experience has been it really does seem random, especially with science ultron drones in aq.

That's another great idea! Thanks everyone for posting, these were all really great comments!

Darkraw346

So do u ask if the chances to evade stacks on top of each other, or if every chance dosent affect the other?

BadPullsMarco

Yes, I asked everyone’s opinion on those two possibilities but got some great answers that seem to definitely point in the Independent direction. Fun thread, I’m a math teacher (and got one of my degrees in it as well) so I definitely really enjoyed everyone’s opinions!