Statistical Mystical!!? What?

[bastuncriochnatha]

Hi,
Tried something for the craic today. Granted it's a small sample but I am still finding it hard to credit. Can anyone assure me that I am not in the Matrix?

Practice mode:
1st and 2nd fav alternating.
Coin toss determines Bet or Lay.
Once position opened ....closed after 15secs.

Bets varying between £30 and £15 depending on odds

Averaged about 10 positions opened and closed on each market (horse) per race.
All activity took place between 8 min and 30sec from off.

8 race results:

-7.50
-2.64
-0.40
-11.03
-0.32
-9.96
-11.11

..and , finally, 25 min out with 117k matched
-15.82

It does seem to fly in the face of probability

Cheers

[ShaunWhite]

Each market irrespective of number of trades with be 50/50 up or down so 2^7 (or 128/1) that those results matched the first outcome. 2^8 (or 256/1) if you were deliberately targeting 8 wins or 8 losses.

It's unusual (probably happens about once out of 128 tries) but it's not quite lottery levels of freakiness.

[bastuncriochnatha]

Thanks for that, Shaun.
Yes indeed, unusual. The sample size is, of course, statistically very small. Like coin tossing 8 HEADS in a row....a bit odd but not yet highly improbable.
I will, of course, continue the "experiment" and should the phenomenon persist in the face of increasing sample size ( say, 50 coin tosses and only 5 tails), in other words, no hint of a strong reversion towards mean then I will be looking for a "rip" in the Fabric of Reality A bit like finding a long sequence of zeroes in Places of PI, eh?

[poklius]

If you are taking price, you are losing 1tick per trade, account for that too.

[Jukebox]

Opening and closing at best price 15 seconds later can only beat the spread if there is movement - 8 minutes out the movement isn't exactly exciting and neither is the spread - Considering that you've exposed yourself 10 times or so in each event I would say that your results are quite normal.
If there is a coin toss analogy then for me it would be that the coin is already at heads and most of the time its not even getting tossed.

[bastuncriochnatha]

Thanks to Poklius and Jukebox.

Average Stake: £22.50 Average Loss: £7.35

Poklius: Valid point about taking price (which I had to do in order to get position opened quickly). I had considered that the tick loss per trade might have a significant effect. Whether it would have sufficient effect to account for the above Av Loss against Av Stake in a random process over the time period is moot.

Jukebox: Take your point about lack of movement and therefore limited, if any, spread around the 8min mark. I had noticed the "coin already at heads and mostly not gettin tossed" occurence more frequenly there and thereabouts of course but I tended to be mainly operating at times when there was high activity (fav and 2nd fav from about 3 min) and the taken price would usually be met with movement pretty much within seconds. Mostly, may I say, moving against the position!!

I must say that I appreciate your comments and will take them onboard as I try to tighten up the test.

As you will have seen, I have no pretentions towards statistical competence. I did it for the craic expecting a standarised outcome and was gobsmacked at the outcome and so it has piqued my curiosity.

Shaun: I tested the statistical expectation of of 8 recurring heads using https://justflipacoin.com/ ...making an assumption as to it's randomised competence. I had gone past 300 trials without success when I got a grip on myself and went for a pint.