He didn't say contingent bias, you did. Nothing he's written is in any way related to prior individual rolls.

He said he knew how to place a winning bet, he never said it was because the previous results were exactly 1, 2, and 3 and no other numbers. Those results can indicate something other than a "contingent" bias; because they landed on 1, 2, and 3, the wheel is leaning more towards that half of the wheel, meaning he could therefore assume bets on that half of the wheel were more likely to hit, which would then allow him to accurately say he knew what to bet next (a complex bet) that would pay out (win more than wagered).

You're thinking about this mathematically, when the edge comes from mechanical physics.