Thought Torrent - Super Scrabble

Super Scrabble was created recently, so I'd expect the letter distribution to be better. Having double the number of letter tiles (196) makes the best fit different for each dictionary:
Best fit for TWL:
11 tiles: E
10 tiles: S
8 tiles: A, I
7 tiles: N, O, R, T
5 tiles: L
4 tiles: C
3 tiles: D, G, M, P, U
2 tiles: B, H, Y
1 tiles: F, J, K, Q, V, W, X, Z

Best fit for SOWPODS:
11 tiles: E
10 tiles: S
8 tiles: A, I
7 tiles: N, R, T
6 tiles: O
5 tiles: L
4 tiles: C
3 tiles: D, G, H, M, P, U
2 tiles: B, Y
1 tiles: F, J, K, Q, V, W, X, Z

Barring the zero tile rule, J and Q wouldn't have any tiles. X would have one for TWL, but none for SOWPODS.

The Super Scrabble distribution does horrifically using this metric:
Mean squared error versus TWL:
Super Scrabble : 90.7515169087742
Best Fit TWL: 0.143839388321553
Best Fit SOWPODS: 0.180594466944665

MSE versus SOWPODS:
Super Scrabble: 91.4580904851635
Best Fit TWL: 0.165195680628715
Best Fit SOWPODS : 0.151734983822096




It looks like they just took the regular scrabble distribution, doubled it, and then tweaked it a bit. Hrm.