The RNG is broken! (No, really)

I was reading through the FAQ and found the answer to the question &quotI failed an easy challenge five times in a row. Is your random number generator really random?&quot:

[quote]
Leaving aside some highly technical qualifications — yes, it is. Randomness is streaky. There are so many individually unlikely possibilities that, if you play often, you will almost certainly encounter some of them.

If that doesn’t convince you, you can take a look at this scatter graph of ten thousand randomly generated numbers from Sunless Sea. Should you wish to subject it to statistical analysis, we’ve also made the raw data available on Pastebin. We’d be amazed if you find anything out of the ordinary, though (apart from the outline of a wallaby visible on the top left).[/quote]
I looked at the data on pastebin and the graph, and now I’m actually convinced the rng is wrong! (And I don’t think counting to 100 is that technical.)

I assume that these values are (supposed to be) used for storylets with a skill or luck challenge, which always show the chance to succeed as a percentage.

Looking at the graph (or through the pastebin), it shows that random values have been generated from 1 to 99: both 0 and 100 never occur. That’s 99 unique values, just to be clear.

But in order to replicate a random event with a probability of a certain percentage exactly, you need 100 different outcomes, each with a probability of 1%. This would mean that for each random event the probability which is shown is not equal to the actual probability, as each unique value occurs with a probability of 1.01%. This would also mean that there are two values where the probabilities are equal: it could be that a shown chance of &quot1%&quot happens with probability 0%, or &quot99%&quot actually happens 100% of the time! We can’t tell without knowing how these values are used.

Alternatively, these values were not generated by the game itself but by a simulation, which makes me wonder if the results may not correspond to the game in more ways… ;)

They clearly stated that these numbers came from Sunless Sea, not Fallen London. There’s also no indication that they are meant to give results in percentages.

The numbers are there to try and prove that you can find all kinds of &quotimprobable&quot streaks in a randomly generated list of inputs. (the question was specifically about encountering several failures in a row, and their answer is about streaks).

Consider known effect of ego, when people assumes 66% win chance for them as somewhat 95% chance, while 66% chance to lose is treated right.

But streaks of bad luck is a scourge of gamer experience via RNGs indeed! Often encountered solution is an increment of actual chance during every failure in a row with a guaranteed effect after a number of tries. On the other hand, I don’t now if this sticks up with FL / SS gameplay, especially SS, where stories meant to be played &quotas is&quot, with failures forming a story as well as successes.
edited by curtistruffle on 3/7/2018

They clearly stated that these numbers came from Sunless Sea, not Fallen London. There’s also no indication that they are meant to give results in percentages.

The numbers are there to try and prove that you can find all kinds of &quotimprobable&quot streaks in a randomly generated list of inputs. (the question was specifically about encountering several failures in a row, and their answer is about streaks).[/quote]

What does Fallen London have to do with it?

They give the numbers in response to a question about easy challenges and invite us to statistically analyze them. What would be the point of analyzing those numbers if they are not generated for those challenges? I might as well analyze some dice I’m rolling myself, both would not help determine if the random numbers used by the game are actually random.

They clearly stated that these numbers came from Sunless Sea, not Fallen London. There’s also no indication that they are meant to give results in percentages.

The numbers are there to try and prove that you can find all kinds of &quotimprobable&quot streaks in a randomly generated list of inputs. (the question was specifically about encountering several failures in a row, and their answer is about streaks).[/quote]

What does Fallen London have to do with it?

They give the numbers in response to a question about easy challenges and invite us to statistically analyze them. What would be the point of analyzing those numbers if they are not generated for those challenges? I might as well analyze some dice I’m rolling myself, both would not help determine if the random numbers used by the game are actually random.[/quote]
Random numbers are used all the time for lots of things. For example - to roll the range of bundles of oddities, the determine card draws, to determine rewards out of a range, etc.

You assume that these particular numbers are for calculating percentages of success in a challenge.

That is not a wallaby…It is obviously a giraffe