Pilfering Chance Not 80% For Me

[kafka]

I kept track of 100 Pilfer attempts and Missed 38 times. An 80% chance rate should have Missed only ~20 times. I've seen a few reports of similar experiences. Perhaps we're just unlucky. ...Or perhaps someone can check the code.

[小龙马]

I guess you didn't pray to the RNG God beforehand....lol...

On the more serious side, i have had 2 consequtive pilfer misses before. Maybe you will be able to get a "more accurate" amount if you made more attempts. Since it is just a probability, the end results might fluctuates.

Though if you indeed did try that many attempts, the 80% doesn't really matter as it will turns to 90% then 100% when your skill levels up through all the fights you did....unless you try by steal and run.

[Roberto]

... That's now how probability works. Flipping a coin is 50/50 but it won't be a 50% split until you've done it 10's of thousands of times.

[kafka]

Thank you both for not mentioning confirmation bias. ;)

I did track more results and got much nearer to 80% afterwards. In this case, I would suggest that the probability distribution is more oblate than necessary. In other words, the results for an 80% chance should perhaps vary mostly between 70% and 90%, instead of 60% and 100%. ...Certainly not a game-breaker.

[Songbird]

If those original 100 pulls were sampled fairly (i.e. you aren't just estimating from memory), then either there was something biasing the sample (maybe steal isn't actually 80% on every enemy, maybe something weird with how the RNG works), or it was an extraordinary coincidence. The probability of 62 successes in 100 or fewer is only about 0.000025, or 0.0025%, or about 1/40000, if we assume that 0.8 is indeed a constant chance of success.

Were all tests done on the same enemy? Part of me is wondering if the game is running an accuracy check in addition to the steal check, or something like that.

[kafka]

I kept notes on whatever enemy I could steal from until I reached 100 attempts (with the exception of Ghosts, not knowing if their physical defence buff might affect Sienna's chances). I also noted how often I Missed twice in a row, which happened 5 times instead of 4 as the odds might predict, although another double Miss did occur almost immediately after I'd completed the 100 attempts.

Conclusion: I've decided never ever to buy lottery tickets. ;)

[小龙马]

Now now....there is no need to say someone is nonsense, insane or delusional...is not like OP is hurting somebody.

Every one got their own curiosity, things that they like to do etc.

eg. There are those who like to max out every stats even though it is not needed to win the game but they still do it for their enjoyment etc.

[kafka]

One person was concerned that I made my observations from memory. The next person says I was crazy to take notes. You just can't win. lol

[Tenseiken]

Originally posted by 小龙马:

Now now....there is no need to say someone is nonsense, insane or delusional...is not like OP is hurting somebody.

Every one got their own curiosity, things that they like to do etc.

eg. There are those who like to max out every stats even though it is not needed to win the game but they still do it for their enjoyment etc.

To be fair, going around with only anecdotal evidence/very little evidence (like, extremely little, really) in general and throwing out accusations are rather rude towards the very creators of this game, so I wouldn't say it's not NOT hurting somebody. It's pretty unfair.

For some reason I do however want to give you the benefit of the doubt that you're not doing it to be malicious.

Then again I'm just wondering why go these lengths when you get to lvl 3 Pilfer (100% chance) SO fast. Not to mention you could 100% the game and keep Pilfer at lvl 2 for the sake of curiosity (80%) the entire game and technically still not have nearly enough data to show.

Such a weird thing to be alerted about considering how little time you spend on lvl 2 with basic knowledge of probability and chance. I don't even mean that in an offensive way as much as it sounds like it. I'm just genuinly shook how and why.

But as I said you sound like you have the well-being of the game in mind so I will discontinue my angry parent larping.

[Tegridy Farms]

100 times just isn't enough test to factor out variance clumping. You would need to do atleast 10 thousand test and get a result somewhere around + or - 5% to get a large enough sample size to really convince someone who spends a lot of time analyzing statistics there is an anomaly and even then they would probably want to run a few extra sets of 10 thousand and see what happens. All a sample size of 100 tells you is you tried 100 times and missed 38 you could very well keep track of another set and land all 100 of them.

[kafka]

Originally posted by Tenseiken:

...throwing out accusations...

What accusations are you referring to?

Please re-read my first post where I wrote, "Perhaps we're just unlucky."

Then re-read my second post where I wrote, "I did track more results and got much nearer to 80% afterwards." and "...Certainly not a game-breaker."

Obviously, I was trying to help improve this great game. Not obvious enough for some.

[kafka]

Originally posted by Tegridy Farms:

...All a sample size of 100 tells you is you tried 100 times and missed 38 you could very well keep track of another set and land all 100 of them.

As I mentioned above, the probability distribution might be more oblate than necessary.

[Shokudaikiri Mitsutada]

Damn the only hard part about stealing enemies is when they are bosses. I realized that it cost me like 3 or 5 turns to steal an item from a boss. Normal fighting is easy since I have a 90% chance to steal with Sienna plus a 50% chance to steal an extra item with a passive I equipped on her.