UK Lotto 659 Test 10 - Sat 6 May 2023

[NN]

RAND : [32, 24, 26, 29, 47, 49, 55, 53, 22, 34, 9, 51, 11, 4, 8, 7, 36, 3, 23, 16, 48, 12, 42, 40]
CONST : [4, 23, 9, 37, 49, 55, 14, 40, 20, 7, 34, 1, 45, 22, 57, 17, 44, 56, 15, 16, 21, 32, 54, 31]
FIB : [4, 58, 19, 54, 3, 43, 45, 55, 2, 21, 37, 11, 28, 23, 38, 15, 24, 31, 39, 34, 27, 56, 42, 6]
PRIME : [31, 16, 46, 23, 19, 12, 25, 41, 58, 56, 6, 48, 42, 10, 29, 52, 45, 24, 17, 34, 22, 39, 38, 15]
CONTROL : [14, 48, 22, 18, 26, 29, 15, 33, 41, 59, 21, 50, 28, 39, 34, 25, 38, 23, 54, 58, 19, 12, 3, 49]

[NN]

Result = [3, 23, 33, 36, 54, 55] + bonus= 7


RAND caught: {3, 36, 23, 55} +b ==> score: 39 [18, 8, 7, 6]
CONST caught: {23, 54, 55} +b ==> score: 44 [23, 19, 2]
FIB caught: {3, 23, 54, 55} ==> score: 69 [21, 20, 17, 11]
PRIME caught: {23} ==> score: 21 [21]
CONTROL caught: {33, 3, 54, 23} ==> score: 32 [17, 7, 6, 2]
Nan caught: {36} ==> score: 18 [18]
Leela caught: {33, 3} ==> score: 33 [18, 15]
Alphazero caught: {33, 36, 23} ==> score: 36 [18, 14, 4]
--------------------



fib [37, 46, 1, 31, 19, 50, 5, 42, 9, 69] = 309
const [44, 18, 32, 32, 47, 15, 13, 50, 9, 44] = 304
prime [44, 18, 8, 31, 14, 13, 31, 52, 37, 21] = 269
rand [0, 0, 43, 8, 16, 51, 12, 48, 33, 39] = 250
nan [5, 5, 20, 49, 31, 34, 32, 21, 14, 18] = 229
control [0, 0, 37, 28, 11, 10, 28, 60, 6, 32] = 212
alphazero [0, 0, 37, 28, 11, 10, 17, 22, 16, 36] = 177
leela [0, 0, 37, 28, 11, 10, 13, 31, 0, 33] = 163

[NN]

Fib and Prime have done better than I imagined.

If the 1st and 2nd results are ignored then fib and prime are closer to 'control'
leading me to conclude

(1)Its better to space out the picks

(2) I tried to pick roughly 50:50 = odd:even , although on occasions 11-13 or 14-10
were allowed ( 24 numbers)

(3) I tried to pick roughly equal numbers in every decade.
( which was only true for const only )
for the others you would sometimes get 5 or 6 in the same decade and less in others



example

RAND : [11, 29, 26, 4, 49, 19, 17, 33, 35, 30, 22, 44, 10, 2, 34, 20, 6, 43, 8, 12, 31, 55, 24, 28]
CONST : [6, 27, 47, 54, 18, 11, 4, 42, 32, 31, 35, 55, 37, 58, 19, 49, 9, 29, 2, 24, 16, 44, 53, 22]
FIB : [38, 7, 16, 41, 40, 11, 55, 58, 52, 36, 27, 20, 6, 30, 3, 54, 59, 28, 46, 47, 22, 24, 21, 17]
PRIME : [14, 44, 4, 38, 27, 21, 45, 51, 57, 12, 53, 3, 18, 2, 15, 20, 41, 28, 52, 5, 9, 55, 23, 50]
CONTROL : [13, 20, 41, 50, 35, 39, 52, 36, 5, 10, 2, 46, 54, 29, 16, 1, 34, 23, 45, 40, 51, 14, 4, 12]
------------
(10, 14) (4, 5, 6, 5, 3, 1)
(12, 12) (4, 4, 4, 4, 4, 4)
(10, 14) (3, 3, 6, 3, 4, 5)
(13, 11) (5, 4, 5, 1, 3, 6)
(10, 14) (4, 5, 3, 4, 4, 4)

(odd,even) (1-9,10-19,20-29,30-39,40-49,50-59)


Hope someone finds this useful...

[Marmaduke Jinks]

Post by NN
Fib and Prime have done better than I imagined.
If the 1st and 2nd results are ignored then fib and prime are closer to 'control'
leading me to conclude
(1)Its better to space out the picks
(2) I tried to pick roughly 50:50 = odd:even , although on occasions 11-13 or 14-10
were allowed ( 24 numbers)
(3) I tried to pick roughly equal numbers in every decade.
( which was only true for const only )
for the others you would sometimes get 5 or 6 in the same decade and less in others
example
RAND : [11, 29, 26, 4, 49, 19, 17, 33, 35, 30, 22, 44, 10, 2, 34, 20,
6, 43, 8, 12, 31, 55, 24, 28]
CONST : [6, 27, 47, 54, 18, 11, 4, 42, 32, 31, 35, 55, 37, 58, 19, 49,
9, 29, 2, 24, 16, 44, 53, 22]
FIB : [38, 7, 16, 41, 40, 11, 55, 58, 52, 36, 27, 20, 6, 30, 3, 54,
59, 28, 46, 47, 22, 24, 21, 17]
PRIME : [14, 44, 4, 38, 27, 21, 45, 51, 57, 12, 53, 3, 18, 2, 15, 20,
41, 28, 52, 5, 9, 55, 23, 50]
CONTROL : [13, 20, 41, 50, 35, 39, 52, 36, 5, 10, 2, 46, 54, 29, 16, 1,
34, 23, 45, 40, 51, 14, 4, 12]
------------
(10, 14) (4, 5, 6, 5, 3, 1)
(12, 12) (4, 4, 4, 4, 4, 4)
(10, 14) (3, 3, 6, 3, 4, 5)
(13, 11) (5, 4, 5, 1, 3, 6)
(10, 14) (4, 5, 3, 4, 4, 4)
(odd,even) (1-9,10-19,20-29,30-39,40-49,50-59)
Hope someone finds this useful...

I find it useful. Thanks :-)

I've always liked Fibs and Primes as a "set".

In a spreadsheet I've got 7 sets of 10 lines. In each line, within the set
of 10, I have 2 PrimeFibs; along with at least two numbers divisible wholly
by 3 and a Family Group of numbers (like 1 and 11 or 3 and 33) and a number
wholly divisible by 5.

Much to my chagrin it's not produced anything huge. It's just for fun ;-)

Jinks

[nigel]

Post by NN
(1)Its better to space out the picks

Positional limits!

Nigel