Remember when I said I got it twice in one stream?
Something changed before that happened that I didn't consider relevant at the time, but I do now. So I'll post the data and let you decide:
24th December, 1986: 3200 attempts
S-prize: 0 (0%, between 0 and 1/1066)
1 prize: 30 (0.94% (~1/107), between 1/168 and 1/78)
2 prize: 62 (1.94% (~1/51), between 1/68 and 1/41)
3 prize: 96 (3.00% (~1/33), between 1/42 and 1/27)
4 prize: 195 (6.09% (~1/16), between 1/20 and 1/14)
5 prize: 408 (12.75% (~1/8), between 1/9 and 1/7)
No prize: 2409 (75.28% (~3/4), between 73.75% and 76.81%)
25th December, 1986: 1600 attempts
S-prize: 4 (0.25% (1/400), between ? and 1/200)
1 prize: 15 (0.94% (~1/107), between ? and 1/70)
2 prize: 30 (1.88% (~1/53), between 1/84 and 1/39)
3 prize: 41 (2.56% (~1/39), between 1/57 and 1/29)
4 prize: 102 (6.38% (~1/16), between 1/20 and 1/13)
5 prize: 197 (12.31% (~1/8), between 1/10 and 1/7)
No prize: 1211 (75.69% (~3/4), between 73.54% and 77.83%)
For reference, the official odds:
S-prize: 1/400
1 prize: 1/100
2 prize: 1/50
3 prize: 1/10
4 prize: 1/8
5 prize: 1/5
While 3/4/5 prizes are still definitely wrong, I have to say that my attempts on Christmas Day look a lot closer to the official odds than the Christmas Eve attempts. Of course, this sample size is far too small to make any concrete claims - it's still very much within the realm of random chance; but it does seem as though Christmas Day is a "lucky day" in a way that Christmas Eve isn't. I'll be doing more testing today and trying to figure out exactly what's going on.