Would you trade a coin toss?

[Effkay]

Surely using a coin toss as a basis for trading decisions is a mugs game looking at the probability, since the coin and the market are unrelated.

What I mean is presuming u toss a head (meaning go long)

thats a 1/2 chance

The market has a 1/2 chance of going up( simplistic view!),

so the chance of you making the correct trade is 1/2 * 1/2 = 1/4

So from a probability point of view you only have a 1/4 chance of success?

Anyone agree? Or should I shut up now?!!

 

[LION63]

Effkay,

You are right as the two events are unrelated, as such, the chances of the two occurring together are 1 in 4. Not a clever way to make money or preserve one's capital.

 

[trendie]

Surely using a coin toss as a basis for trading decisions is a mugs game looking at the probability, since the coin and the market are unrelated.

What I mean is presuming u toss a head (meaning go long)

thats a 1/2 chance

The market has a 1/2 chance of going up( simplistic view!),

so the chance of you making the correct trade is 1/2 * 1/2 = 1/4

So from a probability point of view you only have a 1/4 chance of success?

Anyone agree? Or should I shut up now?!!

This has got to be the single best argument against coin-flips.
I never thought about it this way !

The data has to come from the market itself. Enlightenment.

 

[TheBramble]

Surely using a coin toss as a basis for trading decisions is a mugs game looking at the probability, since the coin and the market are unrelated.

What I mean is presuming u toss a head (meaning go long)

thats a 1/2 chance

The market has a 1/2 chance of going up( simplistic view!),

so the chance of you making the correct trade is 1/2 * 1/2 = 1/4

So from a probability point of view you only have a 1/4 chance of success?

Anyone agree? Or should I shut up now?!!

Not really, no.

Chance of a head (or tail) 1/2.

Chance of stock going up (or down) 1/2 - ignoring non-trending [which you shouldn't!!!]

So regardless of what you get - head or a tail - the chance of the market going in your randomly selected direction (ignoring non-trending conditions) is 1/2 - not 1/4.

 

[trendie]

Coin-flip: Heads / Tails.
Market: Up / Down.

Possibilities:
Heads / Up
Heads / Down
Tails / Up
Tails / Down

The chances of market going up or down is, indeed, 1/2.
However, the chances of the coin giving you the correct direction, is 1/2.
So, it must be 1/4. No ?

Anybody want a second-hand coin ?

Dave, this is causing my logic circuits to overheat.
Dave, this is causing my software to melt.
Dave... Daisy, Daisy, Give me you answer do.........

 

[Effkay]

Yes but both have to occur for the trade to be successful

if you get heads and the market goes down it's pointless, and there is a 1/4 chance of that happening
if you get heads and the ,arket goes up thas good 1/4
if u get tails market down 1/4 win
if you get tails market up 1/4 loss

so I suppose there are 2 good outcome and 2 bad ones bringing us back to 1/2 chance of a successful trade?!!!

 

[Effkay]

woops sorry trendie didn't see your post.....but yeah i agree!

 

[TheBramble]

Coin-flip: Heads / Tails.
Market: Up / Down.

Possibilities:
Heads / Up
Heads / Down
Tails / Up
Tails / Down

The chances of market going up or down is, indeed, 1/2.
However, the chances of the coin giving you the correct direction, is 1/2.
So, it must be 1/4. No ?

Anybody want a second-hand coin ?

Dave, this is causing my logic circuits to overheat.
Dave, this is causing my software to melt.
Dave... Daisy, Daisy, Give me you answer do.........

Aaaarrrghhhhh.

ILTWTL.

If you're allowing the coin toss to direct your subsequent market direction for your chosen instrument it doesn't matter what comes up - head or tails. Heads you go Long; Tails you go Short.

There's a 1/2 chance of your selection being correct (normal caveat on non-trending 70% of the time excepted).

So, market direction selected - instrument either does OR doesn't go in 'expected' direction. 1/2.

"I'm half crazy all for the love of you..." NAAARRTTTT!!!!!!

 

[peto]

So, market direction selected - instrument either does OR doesn't go in 'expected' direction. 1/2.

YES!

You will definitely 100% get either a head or a tail, just don't know which. So you will enter the market. It becomes 50% now which way it goes.

The trouble with this method of trading is when you then apply Stops (actual or mental). If you don't you are mad (exposed to unlimited losses); but as soon as you do, you reduce your chances of closing the trade at a profit to below 50%.

Pete

 

[Effkay]

A way of looking at it could be:

The market has already decided without anyone knowing whether it's going up or down, so if it has decided to go up, there is a 1/2 chance of you flipping the coin correctly, and vice versa. So the probability stands at 1/2 of you being correct.

Anyone?

 

[smullaney]

Interesting stuff...here is my coin tossing forex trading for the last couple of months. I should stress this is paper only Curiosity got me to try this out. Basically flip a coin each morning around 9 or 10am GMT and close out at end of day (or stopped out).

Three worksheets - 30, 50 and 100 point stops. Euro/$ and GBP/$, assuming 5 point spread and no slippage. Heads = buy, Tails = sell

I stopped the 30 point investigation as it was going nowhere in a hurry.

Not yet sure if I will put my real money there yet!

 

[peto]

Interesting stuff...here is my coin tossing forex trading for the last couple of months. I should stress this is paper only Curiosity got me to try this out. Basically flip a coin each morning around 9 or 10am GMT and close out at end of day (or stopped out).

Three worksheets - 30, 50 and 100 point stops. Euro/$ and GBP/$, assuming 5 point spread and no slippage. Heads = buy, Tails = sell

I stopped the 30 point investigation as it was going nowhere in a hurry.

Not yet sure if I will put my real money there yet!

Prove me wrong then Smullaney!

30 point stop, 48 wins out of 110 trades = 44% wins

50 point stop, 66 wins out of 102 trades = 57% wins

100 point stop, 24 wins out of 36 trades = 77% wins



You had a bit of luck with the £/$ 100stop: you had 11 wins in a row (1/1000 chance of that happening). It kind of skewed the results on a sample of only 36 trades for that stop level.

As for the other stop levels, I'm suprised the results are so good, in view of the 5 point spread.

interesting

pete

 

[peto]

hi sm

spotted a mistake on 100pt table, line3 column N should read 78 point Loss ( not 68 profit)?

Hopefully just a one-off I ain't checking the lot!

pete

 

[peto]

I should stress this is paper only Curiosity got me to try this out. Basically flip a coin each morning around 9 or 10am GMT and close out at end of day (or stopped out).

sm

Just had a nasty thought about this....

re. my post 39 on this thread..... I suspect you may not have taken into account those days when the price passed through your stop, and then retraced itself to look like a profit at the end of the day????????????????

If so then your losses will be much larger than shown, because in real live trading those stops would have been hit. The profits shown on those days would not exist..........

sorry

pete

 

[smullaney]

If there is a mistake in the figures then it is an honest one, thanks for the catch though. Just fixed it up in both the 100p and 50p charts - absolutely no idea what I was doing there. Anyway, I fixed up the tables.
So as of last trade:
50p stop: Euro/$=174, GBP/$=547
100p stop: Euro/$=122, GBP/$=421.

I will go through them again to double check.

As for the comment about my lucky run, couldn't agree more, damn lucky run!! But isn't that the point of tossing a coin, luck? I am trying to prove to myself that it is money management, not the technique that gets you there in the end. May be prooved wrong, who knows - time will tell.

As for stops, yes I did take account of being stopped out - if the price went through my stop it was closed at the stop price. That is why I have the open trading time there. The next morning I review the day and check the stop was not breached after my opening time. So Pete, no problems there ;-)

I personally didn't think the results are that great myself - you see these "gurus" touting on their websites 1000 points a month....

Cheers
Stephen

 

[Effkay]

I've got a plan....let's see how this goes

The idea is that we want to keep the coin central to the theme, yet have some sort of other 'edge'.

Ok...at 7am (UK) flip a coin, note it, if the next 5 min bar breaches in the direction you want to go, i.e. if the high of the 6.55-7am bar is 1.3356, and you have heads, then only buy if you have a breach of 1.3356 between 7 and 7.05, and the opposite for the short.

if there is no trade flip again.Keep a 20pt stop and a 40pt take profit. stay in the trade until knocked out and flip again. Trade until 9pm. Close all trades.

This still has randomness inbuilt, as the coin gives you direction, and the breach confirms it.

See how it goes.

At the moment I am testing 4 forex systems,
1)EMA crossing
2) if the date is odd/even
3)Coin toss
4)alternating buy/sell

The idea behind all of them is to see if very simple (even silly) entry can make a profit. I am testing them on eur/us, and gbp/us

As for the coin toss, I have tossed a coin 20times, and am now going to repeat the results.

Good luck

 

[Effkay]

to make life easier for the above plan, here are the coin tosses, trade them in this order, and if you run out, start again.

Tails, Heads, Heads, Heads, Tails, Tails, Heads, Tails, Heads, Heads, Tails, Tails, Tails, Tails, Heads, Heads, Tails, Heads, Heads, Heads

 

[trendie]

to make life easier for the above plan, here are the coin tosses, trade them in this order, and if you run out, start again.

Tails, Heads, Heads, Heads, Tails, Tails, Heads, Tails, Heads, Heads, Tails, Tails, Tails, Tails, Heads, Heads, Tails, Heads, Heads, Heads

Ahhh, deja-vu !!

Thread: Chart of 10,000 coin tosses - post #58

good luck Effkay, on your experiment.

 

[blackcab]

Coin tosses are being used in two different ways here - random entries (tossing a coin to make random entries to a market that itself may not be random) and random market (tossing a coin to generate the market and trading that market using your best, non-random entries/exits). Don't confuse the two - they're completely different.

 

[peto]

So Pete, no problems there ;-)

I personally didn't think the results are that great myself - you see these "gurus" touting on their websites 1000 points a month....

Hi sm

Great, that makes the results interesting again!. There's 251 trades there so the likelihood of that 1/1000 chance of 11 heads in a row coming up is only 1/4, not that surprising. My point was merely that it rather mashed up the results for the 100 point stop, when the sample size was only 39.

Useful reminder though that your money management needs to be up to the job.....gotta be able to live with 11 losses (or more) on the trot too!

It is interesting that you at least broke-even, and in fact kept your head above water. If the markets were truly random then I'd have thought that the 5 point spreads would have got you.
5 points * 251 trades = 1000 point loss overall, would have been a likely result. Statistical anomaly or significant result? And if it's significant then a little further tweaking might move us into profit? ...cue Efkay

As for the guru's 1000 points a month, if they could really do that they'd be far too busy getting the hang of their yachts in the Bahamas to bother with setting up websites and coaxing a measly grand or 2 out of us punters.

Efkay, good, let us know how it goes.

pete