Hey guys,

there have been a lot of threads about the christmas boxes. There have been a lot of bitter tears for people who spent tons of money on boxes without receiving the tier 8 premium tanks as well as people who bought only a few boxes and got amazed by getting a tier 8 premium tank with minimal invest.

In the other thread http://forum.worldoftanks.eu/index.php?/topic/655604-break-down-on-boxes-150-boxes-opened/ Ph3lan said that the probability of getting a tier 8 prem tank is constant per box and doesn't change with time (e.g. on day 1 the probability was the same as on day 1+n).

He also said he didn't wanna disclose the exact probability per box of getting any premium tier 8 tank.

So here is my suggestion: Why don't WE find out ourselves? All we need to do is to get a big enough data sample which size allows as to approximate the dropping probabilitties. We could also do a significance test to ensure the statistical hypothesis is within a secure margin. Also we could do other statistical tests to check for the correctness of assumed probability density function.

All YOU need to do is post the amount of boxes you bought (e.g. 3 + 11 +75) and exactly how ofter you got a tier 8 premium tank within your sample size.

I would do the math and present you my figures after enough sample data has been provided!

This would give people a more secure assessment of how much money they wanna invest because they could calculate the chance of tier 8 prem per money spent!

What do you think? Who's gonna take part?

I'll start:

Boxes: 3 + 3 + 25

Drops: 0 + 0 + 1 (Type 59)

So here are my preliminary **Results**:

Taking Binomial-Distribution as the probability function I've tested the results for a alpha (significance level) of 0.01 or 1% and fitted the data so that the independed variable would be the success probability p and found that the probability of getting a tier VIII premium per crate is **3.5% +- 0.1%.**

This is a similar result to the one linearly calculated in that reddit-thread someone mentioned here.

You have to note that the approximation through this method is far more stable as a basis for calculation than taking linear approximation (number of successfull draws divided by total number of tries) against single spikes/off results.

So lets see how the probabilities are for the Bundle sizes of 3, 11, 25, 75 and 2x75=150 boxes:

To summerize the result:

The probability of about 3.5% to getting a tier 8 prem per box seems plausible. Don't forget that there might be a correlation between outlined results and posting on forum about that bad/super good luck - result.

All in all I consider it a quite fair chance and would have thought that the probability is much lower by WG but I guess their business calculations yielded the maximum amount of revenue setting the probability to this figure.

Any additional thoughts? Something unclear? Any suggestions for more calculations (like probability per euro gained or similar)?

Just let me know! Have fun and good luck!

**Edited by ortega456, 26 December 2017 - 02:34 PM.**