Testing Gambler's Fallacy - part 2

[falkor]

Monitor the 3 Even Chance sections - R/B, O/E, H/L - appearing in a row for 10,000 spins.

1. Start with a bankroll of 0 and track any EC that appears 2 times in a row before placing a single flat-bet of 1 unit on the opposite. Note the final bankroll after the 10,000 spins.
2. Repeat step 1 but track anything appearing 3 times in a row.
3. Repeat step 1 again and again for anything appearing 4, 5, 6, 7, 8, 9, 10 times in a row (and anything above that - sometimes up to 15 or 20).

Repeat the above 3 steps for 100K and 1 million spins (3 different ranges in total). I think it would be better to use the same spin data for all the tests.

At what point in any of those tests does the bankroll end up as a positive? Or does the bankroll always end up in a similar negative balance reflecting the house advantage?
This is not only to test Gambler's Fallacy, but also to test if flat-betting can be as effective as progressive betting. The tests could also be expanded further for flat-betting more than once after the trigger.

[rouletteKEY]

Giving this a full 20 seconds of thought if I had to create a method for betting the opposite on an EC after 2 hits for one spin

I would probably triple or quadruple my bet after each loss and see whether I wanted to combine all wagers into one string of bets or keep everything separate (high/ low running a separate progression from say red / black) and then I would drop back maybe three stages after a win and reboot from 1 once I was in plus (don't know where I would place a stop-loss on a run of losses and wait for a virtual win to re-enter)

But I would break out ALL the possible EC's I could reasonably manage
double streets (natural and the in-betweens)
streets
column segments
wheel sectors, etc

Theory being if I could break the table and also the wheel into enough different segments I could try to limit massive variance and without massive variance (and with a stop loss in there somewhere) the progression would work.

I don't play EC's so its my three minute theory now...but I think that's where I would start

[falkor]

Tested 3 Double Streets over 100K spins:
16 times in a row x 1
15 times in a row x 2
14 times in a row x 2
13 times in a row x 5
12 times in a row x 9
11 times in a row x 20

Going to analyse the 9 sets of 12 in a row next...

[falkor]

9
32
9
18
19
17
19
17
17
32
29
9 (Spin 8259)
5 (Lose)
25 (Win)

9
32
29
20
5
19
32
16
31
30
7
9 (Spin 47839)
20 (Lose)
21 (Lose)

32
6
19
7
16
28
5
20
30
17
31
32 (Spin 51838)
19 (Lose)
30 (Lose)

20
9
7
9
28
33
9
29
32
6
16
30 (Spin 54150)
29 (Lose)
15 (Win)

16
19
28
28
18
18
31
28
32
8
20
16 (Spin 64077)
17 (Lose)
30 (Lose)

29
7
21
19
0
30
28
30
31
5
8
7
32 (Spin 78220)
34 (Win)
25 (Win)

7
18
28
4
5
20
28
33
4
31
9
7 (Spin 79960)
25 (Win)
19 (Lose)

29
31
9
16
8
29
8
8
20
33
5
7 (Spin 88326)
3 (Win)
12 (Win)

8
33
20
32
4
30
30
4
28
28
32
5 (Spin 93336)
3 (Win)
11 (Win)

Flat Bet x 1
-1

Flat Bet x 2
0

Progression - 2 steps
-3

[falkor]

will attach results

[falkor]

will attach results

[falkor]

will attach results

[falkor]

will attach results

[falkor]

will attach results

[falkor]

will attach results

[falkor]

There's 1 bug with zero, but apart from that it seems to be working pretty well! 
I will attach them all as they take up too much space on here.

[falkor]

Results here for 100K - 16 in a row down to 8 in a row.

[falkor]

Some VERY interesting results here! 7-10 in a row are all coming back with a profit for a 2 step progressions @ 100K spins! Yet, 1-6 and 11-16 are negative!

Going to test 7-10 over 1 million spins next @ 2 steps.

[RouletteGhost]

Yes you will find that 1 double street from each dozen is the best EC bet. People will disagree and that is fine. Over 1 million spins you may see a long losing streak BUT the long losing streaks are few and far between and not as frequent as the other ECs

[falkor]

They all got killed off at 1 million, so it seems to the only sure fire way to beat this is to see a progression through till the end - but allow for plenty of extra steps in case of really bad luck!