Thanks Palestis, did a quick test using sequential single table spins from spielbank Duisburg. 1u bets, progression for new BR high on any hit (highest it saw was 4u x 8 numbers = 32u total bet).
Let it run for 54 spins to develop a relatively large imbalance for two sets of Finals, began betting at spin 55:
SPIN # LANDED #
1 4
2 25
3 34
4 8
5 14
6 12
7 32
8 28
9 34
10 14
11 29
12 9
13 31
14 24
15 18
16 2
17 32
18 10
19 14
20 13
21 26
22 21
23 24
24 14
25 31
26 19
27 8
28 27
29 35
30 12
31 4
32 22
33 9
34 33
35 11
36 8
37 17
38 35
39 9
40 8
41 1
42 13
43 3
44 28
45 29
46 35
47 17
48 22
49 30
50 8
51 2
52 22
53 13
54 16
55 23 <---Begin betting lowest two Final groups: 0,10,20,30,6,16,26,36
56 34
57 3
58 17
59 5
60 20 Win
61 31
62 7
63 13
64 0 Win
65 28
66 21
67 26 Win
68 25
69 2
70 29
71 21
72 16 Win
73 30 Win
74 1
75 3
76 11
77 18
78 10 Win
79 4 <---Switch to lowest two Finals, new bets: 6,16,26,36,7,17,27
80 15
81 31
82 11
83 26 Win
84 27 Win
85 7 Win
86 3 <---Switch to lowest two Finals, new bets: 0,10,20,30,6,16,26,36
87 13
88 32
89 28
90 21
91 36 Win
92 17
93 12
94 28
95 1
96 22
97 23
98 29
99 16 Win
Terminate session, all Finals groups are relatively close in total appearances at this point.
NET: +212u
Worked well, even playing it relatively mechanically, without subjectively evaluating the mini-clusters for bet placing (I basically bet them all as soon as the imbalances justified it and stayed on them).
At spin 79, a third low Finals group replaced one of the initial two low groups, I switched it in, then once it got ahead, switched back again to the lowest two.
Looks good.