What's the probability that a 40% success rate fails 18 times in a row? Or maybe if you like, a 60% failure rate to prevail for 18 times in a row.
This is how many times I failed to pull my CLS alive out of p1 of haat. It happened yesterday on my alt account where CLS failed to escape 10 times in a row until grievous enraged and then today my CLS on my other account failed 8 times in a row before grievous enraged.
I am a generaly suspicious person and I suspect that there is no true 40% rate of escape. That's just my opinion tho. Still 18 failures in a row is pretty insane to think about... Worse even the streak isn't over yet, I'll try to see if it continues on the next tank.
0
Replies
Two: 40% chance IS "40% of the time" - that's exactly what law of large numbers say. If you do high enough number of retries of independent events, than the larger the number of retries, the closer the number of successes/total will be to the probability of success in single try: if you have 40% for success than you can have i.e 2 out of 10, 48 out of 100, 3920 out of 10000 etc. Variance can be either way, but it will decrease as the number of retries increases.
well, maybe that's your problem ^_^
I have a very high suspicion that the retries are made to be exactly not independent. They boost streaks, either winning or losing, so that on average over a long time, it's still 40%, but in a shorter series, it's very different than that. Which is what the OP experienced (a losing streak). Next week, he'll have a winning streak (perhaps on pulling full characters out of bronzium cards, or Tie Fighter dodging a ridiculous amount of attacks).
You sound like a gambler who believes in hot streaks
I would be very surprised if that happened 20 times. Probably just like OP. And just like OP I'd suspect there's something wrong with the dice. Apparently he did try another one (account) and his suspicions turned to the game board - obviously
https://swgoh.gg/u/xezee/
To go further down that road, I'd bet that a gambling machine that has streaks (say alternate 100 "lucky" draws with 100 "unlucky" ones) is much more addictive than one that uses a true uniform distribution (all draws independently neither lucky nor unlucky), since hot streaks keep the player engaged with "wow, this is hot, I want more", and cold streaks keep the player enraged and swearing "that can't go on like that forever, my luck has to turn".
You can't just calculate the mean over a large sample size to get an accurate picture of what's going on, statistically, either. Would it surprise you to learn that there is a correlation between what gear pieces show up in the shards shop and what refresh your sampling from, what items were purchased in previous refreshes, or what items are currently available for purchase with crystals? I tracked each refresh for ~60 days along with 8 other variables, and correlations like that were beginning to emerge.
I'm not saying that the escape rate isn't 40% over time, but only looking at the expected value to try to get an idea of the underlying mechanics is like trying to bake a cake with only one ingredient.
I agree. But you have to be careful about seeing correlations where there are none also. 60 days is a good start, but probably still not enough. They are pulling data from so many players that how they build their models could be very hard to see from 1 person's data.
Well and if you play the game for 10 year that's 3650 days so for some people that's not a big enough sample infront of the infinity that it could take to prove or disprove a theory. The thing is that I'm not gonna live 100000 years, I wanna see the results now while I'm playing, not in the long term