Balancing dodge/Hit - non existent -

[Neolithic]

I am at 1752 wins in pet pvp battles and still cant figure out how they balance dodging and chance to hit.
IMHO i think its broken.
Some days it seems like i miss 2-3 times PER battle regardless of the pet combo i use.
Some days ill miss 2-3 times in a 2 hour grind session.

[Kpb321]

I'm not quite sure what you mean by balancing dodge/hit but I suspect that what you are seeing is just a natural result of human perception and having a large sample set. Statistics tell us that even something that hits 95% of the time can and will miss multiple times in a row it is just rare. Unfortunately when you do something lots and lots of times you do see those rare things occur and we tend to notice and remember those rare things much better than we remember the 100's of normal things.

The odds of flipping a coin and getting heads 5 times in a row is .03% but if you flip a coin enough times you'll still end up doing that at some point.

[Se5s]

I'm not quite sure what you mean by balancing dodge/hit but I suspect that what you are seeing is just a natural result of human perception and having a large sample set. Statistics tell us that even something that hits 95% of the time can and will miss multiple times in a row it is just rare. Unfortunately when you do something lots and lots of times you do see those rare things occur and we tend to notice and remember those rare things much better than we remember the 100's of normal things.

The odds of flipping a coin and getting heads 5 times in a row is .03% but if you flip a coin enough times you'll still end up doing that at some point.

it's easy to quote statistics, it's tiresome to collect data, it's hard to really analyze the data and make a concrete assessment of what's really happening. i've lost entirely too many matches based simply on missing that 95% hit chance move and being critted in response. it happens a lot more frequently than the numbers say they should. i think blizzard took the dodge chance and made it extra miss chance that is not shown. the other side is that maybe some people are just really unlucky.

[Kpb321]

If you think your abilities are missing more than they should be you can check for that but until then I'll stick by my explanation that it's just those rare events happening and human psychology making those memorable. I don't claim to be immune to that either as it sure does sometimes feel like things are missing way to much and I some how manage to loose a fight that should be an easy win.

http://wow.curseforge.com/addons/pet-accuracy-recorder/

[Papazol]

I agree, Blizz RNG for pet battles is realy broken. You dont need to check any statistic, if you know what is Westgard rules in quality control. And if "95% hit" ability misses 3 times one by one, you can say with 99.7% confidence it isn't 95% hit chance.

[Worgenbait]

The percentages are fine. If you miss multiple time, it's just bad luck, the roll of the dice. Last night, I needed one hit to finish off a PvP opponent, missed 6 times in a row and he won. pissed me off, but that's just the luck involved. Granted, my attacks were 80 and 90%, but still the odds said I should of hit once in that string. Just bad luck when you get a mess of them in a row.

[Sanzul]

I agree, Blizz RNG for pet battles is realy broken. You dont need to check any statistic, if you know what is Westgard rules in quality control. And if "95% hit" ability misses 3 times one by one, you can say with 99.7% confidence it isn't 95% hit chance.

You're taking that data out of context. The chance of three given hits in a row missing is indeed quite low, but you have to take into account how many battles you actually fight and how many hits are exchanged in each. You've likely fought more than a thousand pet battles by now, with about two dozen attacks made in each one. It would be strange if you hadn't seen numerous triple misses.

Much more importantly, all your pet battling combined is still just one data point out of many. There are hundreds of thousands of other players who are also pet battling. It would be strange if there weren't a few of those hundreds of thousands who were very unlucky and missed a lot. Perhaps that happened to be you. Since you're only posting here because you already noticed that you miss a lot, the likelihood of you being one of those exceptionally unlucky individuals is irrelevant. It's like winning the lottery; it's unlikely to be you, but it has to be someone, so it's far from strange that someone shows up to claim their winnings. Arguing how unlikely it is that the person who shows up actually won is meaningless, because they wouldn't be there if they hadn't done so already.

[Drudatz]

yes its total out of control - the best example is stll anubis idol against each other with sandstorm up:
one with crush the other with demolish gues who hits more often?
right! demolish with 40% hit compared to crushs 70% hits more often

and with 4095/5000 I surly can say I have seen things happen that shouldnt have...

[Sanzul]

right! demolish with 40% hit compared to crushs 70% hits more often

Do you have any actual collected data to back up that statement?

[Breehit]

the forum repeated my submission, so see below

[Breehit]

Of course, unless a person has a significant number of data points (>30 is required for statistical significance), all comments are simply opinion. Nevertheless, my own experience has led me to avoid anything with less than 100% hit probability, because anything else is bound to miss when I need it most desperately. But that is my own relationship with the RNG gods, lol.

[Madsen]

This topic is so pointless. Any attempt to teach an rng-whiner about statistics is just met by them repeating what they first said. They don't want to be educated, they only want to put blame on anything but themselves.

[Lucero]

I did a small test (about 100 hits) with demolish while sandstorm was up. In the end about 60% demolishs hit the target. Which leads me to believe that sandstorm for whatever reason adds 10% hit chance to demolish instead off reducing it. Getting larger sample sizes is rather boring and takes a long time so i guess we will never know the truth