What about it? The impression I have is that this is viewed negatively by most, but I haven't seen it discussed much. I also had the opinion that it is almost cheating or is at least a very sketchy thing to do, and that it seems almost lazy and undisciplined. But now I'm wondering, aside from the fact that it is obviously against the rules of poker sites, why should this not be used as a tool for learning?

We all use tracking software to learn about our games and our opponent's stats, we all use a HUD in-game, many of us are fine with solvers, but then why is buying hand histories bad?

The way we discuss the game begs this question. We talk about "mass player pool analysis" and "what the field does on average", but it seems to me that such claims are either unfounded, based on anecdotal experience, or based on unreliable sample sizes. The sorts of sample sizes that we typically have in our databases seem unreasonable for drawing firm conclusions in all but the most common spots in poker (in my opinion). For example, to understand how a player in EP responds to a 3bet from CO, we need a sample on how they play facing a 3bet from at least any IP player, if not from CO specifically. We can't just see 1000 hands on this player and assume his fold vs 3bet or his 4bet range stat means anything. If we love our statistics so much and rely heavily on tracking software, it seems funny to reject the acquisition of a meaningful sample size so that we can actually use our statistics.

So is there some compelling reason (other than "it is against the terms & conditions") that this topic is frowned upon? Is it even frowned upon? If so, is it more about respecting the game than anything else? If that is the case, then where is the line?