Randomness of laddermap picking algorhithm

I notifced that I played a shit ton of loki yesterday so I decided to look at all the 20 games I played this ladder rotation (only yesterday and today) and got this:
I ran a Chi squared test on it for equal distribution and got a probability of 11.5%. So the chance to get this distribution of maps from a perfect random generator is 11.5%. Unfortuately I don't know what distribution excel uses for it's chi squared test but binomial or normal distribution should not make a huge difference for the calculated probability.

If I just look at the data from the beginning of the ladder rotation up to the latest loki game I get this:
5/12 games were loki. Chi squared test spits out a 0,12% chance for this to be the result of an equal distribution. I know selecting the dataset like that is not how you do this properly but this still looks suspicious. Note that all those games were played within an hour or so.
So question now is: Am I just very unlucky or does the random generator that picks the map have bad seeding?
Note that this is already cherry picking a dataset becaue I was annoyed and therefore looked into it.

Your samples are too small to make any meaningful conclusion. Can you keep tracking your ladder maps and see if this persists?

Or maybe query the api a bit for ladder games

As expected the excel function is not calculating stuff properly. Could have guessed so because it was lacking documentation. I re did the calculation manually and arrive at a probability of 15% for the whole dataset and a probability of 0.25% for the second binned case.
Regarding the too little data. Per se the chi squared test already takes the size of the dataset into account. But If you look at the perspective that I have played 500 laddergames and pick the worst case with a sample size of 10 that is already multiplying that above 0.25% to 12.5%.
The problem with doing it with more data: It is very easy to get a random generator that gets a small chi squared for large sample sizes. But in between can be messy as hell. There are more sophisticated test for testing random generators, but I haven't looked into that yet.

why do we even have so many maps in the ladder pool if we only play 5-6 anyways. I played approximately 20 Ladder games this rotation and only got 5 different maps. My opinion is that we have too many maps in one rotation. We should have way fewer maps but maybe different maps entirely for each bracket. So 1800+ can have other maps than 1500+.

Having pools that do not communicate with one another introduces needless player alienation. I’d be willing to have something like the top 2 brackets not have some maps in <300 since those brackets are separated, but nothing like 1300-1800 having entirely different maps from 1800+.

FAF relies on map diversity to showcase strategic diversity, and judging by the last survey results this is a widely held community opinion. This is why I’m not interested in lowering map pool size.

Lokie is a good map.