Test of Hit & Run vs Continuous play

[Bayes]

Woods asked me to simulate a fairly simple repeater system, and here are the results.

The System

Start off betting the last 5 numbers on the marquee, with 1 unit on each. If you lose, bet the last 6 numbers again with 1 unit/number, if lose again bet the last 7 numbers. If you lose again, return to betting 5 numbers, but this time staking 2 units per number, if lose again bet the last 6 numbers again with 2 units/number, then if another loss bet a maximum of 7 numbers. So the system is to play a series of mini-games with a minimum of 5 and a maximum of 7 numbers per bet. After each mini-game of 3 bets increase the stake on each number by 1 unit.

When you win, reset to betting 5 numbers if necessary, and decrease the amount staked per number by 1 unit, if necessary. If you reach a new high balance then reset the stakes to 1 unit and the numbers bet to 5.

Hit & Run Vs Continuous

There were 2 tests. Hit & Run skipped 350 spins whenever a 500 unit bank was lost, or a profit of 200 units was made. Whenever either of these events occurred, play recommenced with 1 unit stakes betting on 5 numbers.

The continuous test also reset when a bank was lost or after a 200 unit profit, but there was no skipping of spins.

So the only difference between the tests was that one method of play skipped spins, and the other didn't.

Testing was over a 50,000 spins (from a real wheel). Obviously, for a valid comparison, each test must bet the same number of spins. I ran the hit & run test first to see how many spins were used, then I ran the continuous test over the same number of spins.

Results

Obviously, since the tests used different sets of spins, the difference in the result could be due to chance, so don't come to firm conclusions just yet. I'll be testing against more spins later and will post the results here.

Both results were negative, but the Hit & Run test lost 13,627 units, and the Continuous test lost 17,764. This is a significant difference (more than 23%).

Results files are attached.

[Juiced91]

Well then it can't be the same if you are using different numbers. You should give them both 50 000 betting oppurtunities rather than 50000 spins and i know hit and run will take more spins. So why not use your million live spins and see how long it takes to get to 50 000 betting oppurtunities then use those amount of spins continuously. The betting oppurtunities can be the same using the same spins just the length will be different well that's how i see it.

And then we will get to the conclusion that hit and run is the same as continuous. as you should hit the same patterns in the end

[woods101]

Hi Juiced91,

Could you clarify a bit? If comparing h/r to continuous play then it will always be against different numbers or else the results would most certainly be exactly the same. The system is one in which a bet is placed every spin and so both tests have the same amount of betting oportunities (and naturally the same amount of spins), the only difference is one is played against the same number of spins consecutively, the other is played against the same number of spins but with intervals of 350 spins each time.

[Juiced91]

Okay i can try... It made sense to me when i explained it to my self. Maybe its wrong, maybe you/bayes would explain my errors when you understand it.

What im saying is run the hit and run sequence for 50 000bets whether it take 200 000 spins or 2 000 000 spins. Then use the exact same spins i.e (200 000/2 000 000 spins) and run the continuous play on that . Yes its gonna take less spins to hit 50 000bets but thats also the corresponding factor of playing hit and run vs continuous you gonna get to your bets faster. so if it bets once every 350 spins take 350*50 000 and we get 17 500 000 yes thats 17 million   but we can always lower the amount of bets, but we use the exact spins in the exact same place. Just betting locations will be different. Maybe im just echoing what bayes said, and thus saying the same thing..

Anyway thats my point of view. 

[Bayes]

Well, I just ran the 1M spin file and the results this time actually favoured continuous play! There were 137,000 placed bets and the results were -

Continuous: -291,163
Hit & Run: -298,089

Juiced is right in that eventually you will hit all the patterns whether you skip or not, but I was interested to see whether the results in that first test would continue if I tested over multiple samples of 50,000 spins.

[woods101]

@ Juiced.
Ah got ya. Yep I think that's what bayes has just done on both accounts, but with the first test he used a smaller amount of spins.

@ Bayes- I thought we nearly turned you there! Shall I start heating up my custard then?

[Bayes]

Shall I start heating up my custard then?

I think so, Woods. There is certainly merit in resetting to keep stakes low, as my first test showed, but none for skipping spins. I'm under no illusion that this will change the minds of the believers though.