That time seems high. I got ~7:25 in what I assume are load times (all solid-black frames that occur between levels). I ignored room-to-room transitions. Maybe that's where some of the discrepancy comes from? Couldn't tell you the model of PS2.

I'm not telling you to buy a different console, or calling for PS3 runs to be banned.

There shouldn't be a pay-to-win aspect to speedrunning (e.g. "Drop $300 to reduce your time by a minute"). If we ignored load times entirely, we'd have genuine measure unbiased by hardware. There was an SDA thread discussing simply cutting load times out of Tropical Freeze because "getting a better time" wasn't consistently tied to "having better gameplay," and the responders were all fine with the idea.

Is the only reason we're not doing that because it's too much of a hassle to measure? It's more work, but it gives everyone who owns a copy of the game an equal shake.

The runs definitely shouldn't be banned.

Gameplay should be king, though. If one run's 10 seconds ahead of another, it ought to be because that person played ~10 seconds better--not because their console loads text characters faster or something. Someone with worse gameplay shouldn't have a listed time over a minute ahead of someone who played better; it hasn't happened yet, but with this kind of variance, that situation is inevitable.

Ideally we'd just ignore load times entirely. The only reason we're not doing that is how inconvenient it'd be to calculate, right?

It certainly makes a difference (1-2 seconds on each load?), though the gameplay is great and shouldn't be ignored either. I think it could be included under a different category, to indicate that the PS3 and PS2 times can't be fairly compared to each other.