I've sent an Test Stage run with 5:22, but when i compared side-by-side with the actual wr (TommyeAsY), i've noted he was faster than me. So i played more and got another faster time (5:29) and this time was the real fastest.
Comparison side-by-side with a timer
If you see the timer turning a 1 second, he goes from 0:37 to 1:41 and the miliseconds is not synced with the seconds.
My faster time was 5:04 TommyeAsY got 5:25 My another time was 5:55
I had seen some timer oddity with at least one other submission, and it was my suspicion that an RTA slower time would be faster IGT.
However, I feel that in this case, it's possible (though to what extent I do not know) that RTA could also not be 100% accurate due to lag etc., so it's sorta a case of what's less bad.
As such, I wouldn't really mind tracking both RTA and IGT times (although I would have to manually time all of the submissions so far) due to this development.
As for what the boards should be ultimately sorted by, I'm assuming this means you would want it by RTA, and I would likely agree with you. I would like to hear what other submitters think before I switch priority though (in the meantime, I will be manually timing submissions and editing them in).
By the point of lag, you're true!
So will be the IGT?
Thanks for reply ^^ Also, good luck with manual time edit
Ok, I have fully figured out what's going on after some personal testing.
Typically, centiseconds are based out of 100 (00-99), and this game has centiseconds that only count up to 60 (00-59). The game is counting whole seconds right; else there would be further disparity from RTA than, for example, titoncio's 5.22 having roughly a 5.7x RTA.
So, the game "overcounts" centiseconds until there is 100, only then does the next second go forward. (Therefore, it goes from 00-59, back to 00, and then up to 39 again in the same second)
Therefore, the correct ruling is, when there is a centisecond "overcount" right at the end of the stage, instead of, say, 5.22, I will add 40 to the centisecond column, making 5.62, which is what the game would show if centiseconds were out of 100. (This is going to vary, however, I will have to adjust by either 20, 40, 60, 80, or nothing at all)
As a demonstration, the first second will end at a 40. Second 2 starts from there and ends at 20, second 3 will go back to 0, and so forth.
IGT is still a valid tracking method here, and ILs can go on as normal.
If anyone has any additional input, feel free to post, but I'm reasonably certain there's nothing actually wrong with the timer, thus, no need for RTA.
As a follow-up, all times on the leaderboards have now been adjusted accordingly, if they needed to be adjusted (a few did not need adjustment). I will also make some additional notes in the rules about what to do when submitting.
After clearing leaderboard we got a bunch of innacurate times, like Test Stage WR (it's 4:10 now when it supposed to be retimed to 4:90 etc)...
Will we have legit leaderboard in future eventually? Seriously, the fact of approving Release 5 runs when only 7th Release is allowed looks like a very bad joke, even worse than innacurate timer http://www.speedrun.com/sonic_world_official/Test_Stage#Mission_5