Did 14 radio calls, didn't get a single Hunter, likely at a 0.3% probability, indicating glitch
Tony1
Member Posts: 12
I am on chapter 4, but have not yet completed it. Due to that, I would think that I'd only be able to call Scouts, Bruisers, and Hunters. I assume all the types of survivors have an equal probability of being called. That would mean I'd expect to get a Hunter about 1/3 of the time when I make a radio call. Over the event weekend, I made 14 calls (using 15 phones each time), using all the phones I had ever gained. I got only Scouts and Bruisers. Everyone seems to agree that Hunters are much more useful than Scouts and Bruisers. If the probability of getting a Hunter was 33.33%, then the probability of me not get a single Hunter with 14 calls is 0.3%. Yes, way less than 1 percent. I think my game is bugged which caused me to not get a single hunter, thus essentially wasting 210 phones (all the phones I ever got) because I was looking to get an additional Hunter.
I would like to get my phones reimbursed to my account. This also means that I was deprived of the increased drop rates for Epic and Legendary survivors event when trying to get a Hunter, so I won't have a good chance of getting an Epic or Legendary survivor until the next event. I can provide screenshots of the entire situation and more information if necessary. Thank you for any assistance.
I would like to get my phones reimbursed to my account. This also means that I was deprived of the increased drop rates for Epic and Legendary survivors event when trying to get a Hunter, so I won't have a good chance of getting an Epic or Legendary survivor until the next event. I can provide screenshots of the entire situation and more information if necessary. Thank you for any assistance.
0
Comments
The remaining order depends on what classes are unlocked.
After that is shooter.
Then warrior.
Next, hunter.
Finally assault.
Why? Because the game parallels the show. Yes, hunters are more useful, but since season 2, the survivors rely more on melee skills whenever possible. 1) Bullets are a finite resource and must be conserved. 2) Firearms attract walkers. Handguns are used more often than rifles and semiautomatic assault weapons.
I have kept track of all my pulls since starting the game in early December. I have made 136 calls. I have drawn 39 scouts, 31 bruisers, and 9 hunters.
Shooters, hunters, and assaults were unlocked after the first radio event I participated in, so I did not include them here. Now that they are all unlocked, I am starting to track a new data set.
Alibaba lvl49
LINE me: princealibabwa
Aegis Official FB Page
Again, the random number generator in the game is not very good.
TL/DR : Never use your intuition, or low number of results (like in less than 1000) to judge a random number generator.
The only complaint one can make is that the RNG is too random, and that he/she would have prefered a number generator that is not trully random but try to even out the chances of each results. That may be a valid argument, but it is a game designer choice to make.
Remember: The are lies, damned lies and statistics!
Then again, each roll is generated on it's own. So, we are in the realm of permutation rather than standard percentile distribution.
It could be "statistically even", as you describe it, if I only made 100 radio calls, and I got a Scout 31 times, a Bruiser 33 times, and a Hunter 36 times. Those numbers are pretty much even given that small sample size of 100. It would NOT be "statistically even" if your tracking was based on your entire guild's radio calls and you all had a total of 100,000 radio calls and got a Scout 31% of the time, a Bruiser 33% of the time, and a Hunter 36% of the time. That sample size is so large that any small percentage difference between the results would indicate a statistically significant difference, or a statistically real difference, caused by something real and not just a fluke in sampling. The cause creating the statistically real difference would likely be due to Next Games programming in this hypothetical case.
I wasn't really even using statistics in my original post, I was only using simple math to produce a totally accurate probability given an equal 33.33% chance of calling each class, but statistics has found its way into the discussion. Obviously someone doing 1 radio call and getting a Hunter does not mean that the rate of getting Hunters is 100%, and if someone really does do 30 radio calls and doesn't get a single Hunter, then it is very bad luck if not a glitch. If the game's random number generator is really bad, then that's a different story and should be fixed. If the rates of calling each class are truly the same, then my situation would be very unlikely (less than 0.3% chance of happening), but bound to happen eventually to someone. But with Next Games' record of glitches, it's easy to think that it's caused by a glitch (or maybe a bad random number generator) and not just poor luck.
Every call is a separate probability. 1:3 chance per pull. With a coin toss, there is a 1:2 chance of landing on heads. Flipping a coin 20 times would yield 10 heads, right? Well no. I might flip it 20 times and get 18 heads. You might flip it 20 times and get 7 heads. As the sample size gets larger, the results will approach 50%. It's a permutation, not a simple probability.
Using your coin example, if we were having some fun flipping a coin, and I really wanted heads, and you flipped the coin 14 times and it came up tails every time, I would be likely to call shenanigans and accuse you of using a double-tailed coin, especially if you had a record of being untrustworthy (glitchy in Next Game's case). Flipping a coin 8 times and getting all tails is almost as likely as my no-Hunter situation.
Without actually seeing the algorithm, your guess is probably a bit better than mine. I noted a trend and based my hypothesis on that. You seem to have the mathematical background to explain the apparent problem.
Curious though, am I correct in that we are dealing with a permutation as the number of tries increases?
(Googling a good statistics refresher course...)