Effect of Assassins

[sXs]

Indeed I am. And my issue is that the probability of killing a merchant and killing a captain is not identical, as they should be, in my opinion.

We were talking about this in guild chat.

Random generation is the "shit happens" scenario.

Your hypothesis testing is a test of the probability of a desired result.

"shit happens" is not compatible with a hypothesis test.

What you want then is for every 100 casts to come out the same

34 captains
33 merchants
33 admirals

That is not random.

[Haron]

I obviously don't expect a 33, 33, 34 result. Don't be absurd. But the number of merchants killed were above the threshold required to discard the hypothesis. This is rather basic statistical analysis.

[Dmanwuzhere]

I have found that if you lift your leg and squeeze out a fart while casting it always gets a result and a smell proving that shit happens

[Meliva]

I obviously don't expect a 33, 33, 34 result. Don't be absurd. But the number of merchants killed were above the threshold required to discard the hypothesis. This is rather basic statistical analysis.

Well did you only run the 1 test? It is possible you just happened to have bad luck on getting merchants. I could for example do the same thing and use 100 assassins, and end up with mostly admirals killed. Its just luck of the draw really.

[sXs]

I obviously don't expect a 33, 33, 34 result. Don't be absurd. But the number of merchants killed were above the threshold required to discard the hypothesis. This is rather basic statistical analysis.

But that is what you are testing with your null hypothesis.

How do you test for randomness? The fact the you had to discard your hypothesis is proof the result was random, which is proof the card is working as described.

[Banger]

As someone who has taken heavy assassin attacks twice, I can assure you although being random, it seems as if it hits the one you have the most off. I feel like they take your pool and put it in one of those lotto machines and just start drawing out ones at random.

[sXs]

Statistically your result comes outside you confidence interval which means it is an outlier. Outlier prove a randomness to results. Although they are not the norm, they are evidence in and of themselves.

Statistically, you do not discard a hypothesis because of an outlier result of a single test. It only means that the sample size was probably too small.

[sXs]

I have found that if you lift your leg and squeeze out a fart while casting it always gets a result and a smell proving that shit happens

You need to stay away from Noffins half filled Rum bottles.

[Shadowood]

it seems as if it hits the one you have the most of.

If you have more Merchants than Captains... You are doing it wrong.

[DezNutz]

it seems as if it hits the one you have the most of.

If you have more Merchants than Captains... You are doing it wrong.

Well before they fixed the Captain glitch, Kart was heading down that path. Are you saying he was doing it wrong?