If the player parlays, his expected loss between the two bets is the sum of 1/37 = 0.0270 units from the first bet and an average of (1/37)*36*(1/37) = 0.0263 from the possible second bet for a total of 0.0533 units. The answer has to do with the way the house edge is defined.
The astute reader may wonder why the player should accept a win of 1,300, at a house edge of 4.97%, rather than parlay, when the house edge in single-zero roulette is 2.70%. Thus, I would do that rather than accept a win of 1,275 or less. By parlaying a first win on zero himself, the player can achieve a win for two consecutive zeros of 1,296 to 1.
