• Re: DLC Discussion Thread: Avenged Sevenfold and The Pretty Reckless

    "Angel" and "Lips of an Angel" aren't in the challenge valids. Neither are "Oh My God" and "Thank God For Girls", nor "Red Devil", "Devil's Island", "Runnin with the Devil".

    That would probably be a good challenge theme. But it ain't this week's.
  • Re: How do songs get tiered?

    I think almost any degree of statistical analysis you could throw at Antibodies' actual leaderboard results at expert level would indicate its tier is non-representative of players' ability to maintain their tier-star-average with it.
  • Re: How do songs get tiered?

    Rio is on my top 5 list of absurd undertiering as well (that I also wrote off as 'they must have made the tiering decision here at sub expert').
  • Re: PSA: Stop random disappearances of songs - Avoid rating them as '1' by accident

    Which means you can most likely blame the vocalist.
  • Re: An Early Look at Upcoming Changes to Rivals Mode

    My gut tells me that maybe some of the variables just can't be modeled properly, no matter how powerful your computer. Or maybe it's just "complicated" and not impossible. Calibration figures and lag and squeezing and pausing, those could all be included. If a setting can be set, if a button can be pushed, through it into the model.
    I think part of the reason this couldn't have been done before is because of such 'variables' - the model couldn't accommodate differences resulting from calibration or pausing. That would be the whole point of simulating optimal scores under the 'new rules', after eliminating the loopholes that allow calibration or pausing to generate higher scores. If they fix their scoring model it would be independent of those variables. If a 3.247 second long sustain has a maximum whammiable benefit of 0.198 overdrive bars, and that is now independent of calibration variations (everyone is given the ability to achieve that regardless of their calibration), then in theory the optimal path is calculable now where it wasn't before (when calibration actually introduced a variable that it won't anymore).