I haven't lost once with this system yet. Average winnings so far has been about 30 units per set (max BR maybe 100 units so far). But this is using 2 unit chips.
Start with a virtual spin and note down the number that hits first. Let's say it's 23. First bet is always on zero unless zero hit then the first bet is on 1 - makes it easier to follow.

Keep a note of all the numbers that hit from now on, since we don't ever want to bet on them.

23 - bet on 0
31 - bet on 0,1
9 - bet on 0,1,2
24 - bet on 0,1,2,3
16 - bet on 0,1,2,3,4

You do this until you get your first win. If in profit then stop and start over. If on your first win you are behind then take the chip off the winning number but carry on betting the rest in the series, adding a new chip to an unhit number, but avoid placing a chip on any numbers that have already hit. When the 2nd win comes then evaluate the balance once more. You should only quit on a new high otherwise keep playing.

To be tested: does the system ever lose? Presumably the more repeaters that occur the better our chances of striking big on a sleeper? So perhaps a set should only begin after several numbers have repeated on the marquee? Under normal starting conditions I guess the danger might arise if 18/19 numbers hit, but none of ours, so we freeze selection and never go above covering 18. We then wait for one of those 18 to hit (must happen in the next 7 spins or longer?) then we take a chip off the winning one and proceed with betting 17...16....15....14.....13.... eventually should yield a profit I would imagine before getting down to 3 or so numbers. If there's no profit and we are back down to betting only 1 or 2 numbers then I wonder if a progression might help somewhere along the line?

I had my first challenging set. It still won +14 (using 2 unit chips) and I decided to carry on playing. It went below zero then won +4, so there was 2 times I could have pulled out. The max I was covering was actually 19 numbers (not 18), but could possibly go higher than that depending on how many repeats occur. By the time I got down to 7 numbers I was well defeated, so there needs to be a stop-loss based on, say, the 1st, 2nd or 3rd win (or perhaps a progression could begin when nearing 19 covered numbers). I don't think any starting conditions are going to make much difference.

You could do a +1 progression after seven numbers miss and go to the next
bet.  Also have you considered Clockwise/counterclockwise neighbors for this?

Good concept imo.  (Also welcome back Falcor) 

Thanks! Yeah, it's just a concept at this stage. If we were to consider the starting conditions we could perhaps monitor how many repeats have occurred during the last 10 numbers or 20 numbers and then use that as a trigger to start beckoning the sleepers.

Are u "the" falkor

I got another variation that also seems to work quite well:

After the first spin, whatever number the ball lands on, bet on that number to repeat. As the 2nd, 3rd number hits, bet on those to repeat and keep covering any numbers that have appeared once or more. You need to keep playing this way until you are a) in profit and b) the threshold has past, say, 18 numbers that have repeated once or more. You then switch to betting the remaining 18 sleepers until you are in profit then stop. A few things that need testing/tweaking:
1) What's the optimum time to switch from betting repeaters to sleepers if not on a profit and after 18 numbers?
2) Once in profit (or greater profit) on the sleepers should we stop immediately or carry on and for how long?
3) Should there be any progression? Right  now, it doesn't seem to require it.

Yet another good variation is based on TurboGenius' post under the "basics - what nobody wants to hear" topic:
start betting 0 then 0,1 then 0,1,2... each time you win add a unit to the winner - so you continue to snake across and engulf the board whilst getting fatter at the same time!
Testing: need to know when to stop. Usually I am in profit by 1/2 or two thirds of the way across the board.

I think all 3 of these variations are very fun to play - and they don't seem very risky at all! I haven't lost with either of those yet... 

I coded the 3rd one (based on repeats), but it failed in the long run.

Priyanka has a video that features a strategy similar to what I am trying to achieve in the variations above:
link:s://:.youtube.com/watch?v=Ugu7s-X0qk0

After all, I always thought this game should be played with numbers!