After hearing a statistics professor told his class that some numbers in Keno have better odds than others, I wanted to check his claim. So, I went to the Michigan Lottery website and wrote a script to take all 16 years worth of Keno winning numbers. That's 83072 numbers drawn 22 at a time for a total of 3776 rounds of Keno.
I found that the best number to pick is a tie between 05 and 11. Each one appears 28.9989407% of the time. The worst number to pick is 29, as it appears only 24.7351695% of the time.
But does this actually matter? Maybe a little but not much. If you look at the difference, 05 was picked 161 more times than 29. There were 83072 drawn. 161 extra times isn't that bad. The standard deviation of all percents was 0.78.
So, was the professor right? Technically yes, but not really when it matters. What would be interesting to know is if the Michigan Lottery uses an electronic system or not to draw the numbers. There's been much debate over using electronics.
Frequency | Number
934 29 971 45 971 47 976 63 984 52 995 30 997 08 1005 58 1005 75 1008 60 1009 64 1014 38 1014 41 1015 67 1017 06 1018 46 1018 69 1018 70 1018 76 1020 04 1023 37 1024 79 1024 80 1025 14 1025 50 1026 49 1027 77 1029 12 1030 03 1030 33 1034 61 1035 36 1036 20 1036 40 1037 28 1037 65 1038 43 1039 10 1039 19 1040 13 1040 25 1041 56 1045 07 1045 48 1045 73 1046 16 1046 27 1046 62 1047 02 1048 39 1051 57 1052 17 1052 22 1052 23 1053 53 1053 74 1054 35 1056 59 1057 51 1058 01 1058 31 1059 21 1060 54 1061 72 1063 09 1063 15 1064 44 1065 26 1065 34 1066 32 1069 55 1072 66 1076 24 1077 78 1078 71 1079 42 1086 68 1093 18 1095 05 1095 11