Forum

i demand forza last post.

Quoted from Stew

[…]

If a rocket scientist tells you that your math is wrong, it definately is wrong.

If they explain why, sure… I actually want to know
(Though I doubt he’d be bothered to explain)
Either way, with simple logic you can tell that it’s incredibly likely an account was shared, so even if my maths is whacked out somewhere, doesn’t change much.

Anyway, ETF2L forums are a strange place, I’m out for now.

–Edit

Quoted from for science

Just because I don’t like the method, doesn’t mean that your are necessarily wrong and the other position is correct.

It just means that I don’t think you can draw much confidence in the conclusion from your method :)

Thanks for this, rather than just “lol bad” :)

Didn’t realise anyone ever listened to `why` but sure…

So what you seem to be aiming to do is calculate the probability that given 2 accounts have had the same IP registered against them what is the probability that they are account sharing.

A few things you really need to take account of to do this: size of the userbase in etf2l that has access to that pool of IPs, possibly the average lifetime of `ownership` of an IP (depending on your approach), how many times a `coincident` IP between the two accounts was registered by the admins, how many times it wasn’t, size of the set and frequency of occurence of IPs that the two accounts have used (both coincident and not coincident).

Basically you want P(sharing | n coincident IPs )

Quoted from for science

Didn’t realise anyone ever listened to `why` but sure…

So what you seem to be aiming to do is calculate the probability that given 2 accounts have had the same IP registered against them what is the probability that they are account sharing.

A few things you really need to take account of to do this: size of the userbase in etf2l that has access to that pool of IPs, possibly the average lifetime of `ownership` of an IP (depending on your approach), how many times a `coincident` IP between the two accounts was registered by the admins, how many times it wasn’t, size of the set and frequency of occurence of IPs that the two accounts have used (both coincident and not coincident).

Basically you want P(sharing | n coincident IPs )

If all the Turks on ETF2L are active and on the same ISP, that’s 228 people within the pool… considering the overall size of the pool, I don’t -think- that’d increase the probabilities to anything significant still, so I didn’t bother taking this into account… we’re also dealing with 2 specific players, rather than saying anyone could’ve been sharing the account, so it didn’t seem logical to include this. I’ll take your word for it though

Average lifetime of ownership, yes, I agree. I wanted this kind of information but I don’t think it’s readily available, so had to do without.

The “coincident” IP was only registered once, in whatever size their set was (I think the set size was 250 if I remember correctly what I’ve been told), as far as I am aware.
That leaves 249 instances of the IPs not being coincident.

You’re right, and this was always a rather “rough” bit of stats anyway. I think you really have the ask the question though of “Would any of this change the final result significantly?”. I’m not going to think much into that, but at first impressions, it doesn’t seem so…

^Pretty much what Skyride said

Again though, thanks for explaining, nice to see someone willing to ^^

good read 10/10

who got banned?

Quoted from for science

[…]
Thank you for your exceptionally deep and insightful points on probabilities and statistics.

lmfao