Meanwhile Bisco is 18+ Ex

Uh yeah due to EX's going down as well.

found a bisco today: Probably never use it.

"

Shagsbeard wrote:

found a bisco today:

Probably never use it.

SSF bisco holy shite.

I always find my biscos.

PFFT

SHAGS DOESN'T HAVE A

It isn't mine sure but I do get to plaster it on the forum to pretend to be cool

Thanks I_NO for letting me borrow ur Biscos

"

Nephalim wrote:

20% is a lot if you have 6 sextants, atziri frags, breach, a god tier clear speed build etc etc but for the average player disco's collar is a huge noob trap.

I think so too.

FU Cadiro. Elreon best friend forever.

"

LowBudget780 wrote:

60-100 maps is not enough of a sample size to draw any sort of conclusion

It actually might be. It's surprising how few samples you need to start drawing conclusions on a binomial distribution like this. I don't claim to have done the math, in fact, I'm not really the right one to debate the actual main topic at hand here (I went full casual as usual several weeks ago). But I can attest to how statistics work, and you can't just say it isn't enough samples, this time. It's better to argue that his assumptions were flawed (and they could be).

"Number of map drops above Tier n after running k maps" could be a poisson distribution, a type of binomial. If you count up the short run map drop rate, you can also estimate the probability that your observation was an abberation. Probability of your sample set being an outlier gets extremely small very quickly, even when you're just talking about 100 or so observations.

In other words, a run size of 100 maps might actually be enough to draw some real conclusions. But, I'd argue there are a few uncontrolled sources of error here, and you should run the whole thing at least once more under stricter conditions to see if assumptions hold. Big assumption in treating two pools of rolled maps as equal in value. Not all kills have the same chance of success; and, there's a pretty huge variation between maps in the number of white kills the player will get. So, I think the assumptions could use some work.

Do the test again on a pool of white maps to treat a whole map run as one poisson sample.

This thing doesnt even feel like it works to start with as i ran this all league on harb and said the same thing

In fact ive had more ex drop this league then every other league ive played i think and i dont have any mf and i havnt even been running rare maps (mostly magic cause my char sucks LuL)

I would use it i think even if it doesnt feel like it works just in case if i did have one tho (not on my main chars i dont think tho)