I agree with nsw's thoughts.
I think this kind of question could be a great source of info about analytical thinking about poker and not just security in general. I can't see the original linked question because it's been deleted and my rep is insufficient here, but I think this is a fascinating topic and quite poker-centric.
Is sounds like Toby Booth's thoughts were that the answers would focus more on topics like IP addresses and logon patterns. I agree that this is not poker-specific and shouldn't be here. But there are also other patterns that are relevant which would be poker-specific:
- General betting patterns compared across different accounts
- Game selection and table composition
- Tendencies to enter hands more or less frequently when specific other players are in the hand
- Elements of collusion that become clear only when you have knowledge of hole cards from multiple players
This sounds sort of like a fun big data problem, but where the descriptor set is specific to analytical poker. I think this type of question could be fun here and interesting to the thinking poker crowd.