Chart of 10,000 coin tosses

[trendie]

Hi,

This is a very interesting thread !
I am a beginner at trading and its ideas, so please bear with me if I say anything silly.

re: coin-tosses.
Since the stock market, in reality gains an average of X% per year, could you not load the "heads" with that percentage to get a better approximation ?

For example, if the average gain over the past 30 years, or whatever range, is an average of 10%, then when running the coin-toss experiment, load the generator to produce 110 heads for every 100 tails.

Is this a logical thought ? Can this be done technically ?

regards trendie.

 

[dc2000]

Interesting quote " the markets are random " this would infer that brokers could not influence the market by buying or selling

 

[bonsai]

dc2000 - they dont in the long run - but we trade the hard right edge (unfortunately)

trendie - the issue does not revolve around up/down but by how much it moves up or down. So coin tossing is not really going down the right path

bracke - agreed. you need something that will allow for the market to open up or down by at least 250pts or more. You only need one of those to wake you up to the reality of life. But you can get 2 or more in series.
viz-1987- Dont ever think it couldnt happen again.


but perhaps we can make it more interesting ?
I believe the markets show the effects of Quantum mechanics.
so every tick can be the sum of all possible histories !

now, is the next tick up or down ?

 

[growltiger]

Not the toss of a coin

The comparison with coin tosses ( and dies ) cannot be correct. Over an long period of time heads will equal tails, and the die results will be evenly distributed over the die.
Viewing graphs of indices over a long period shows heads win the day ( cotinuing up )

I think that the randoms among us are trying to 'shoehorn' the random approach into the pattern approach, it won't do.

Regards

bracke

Of course the progress of the market is not random (there is now plenty of "inefficient markets" theory that supports the idea that investors can rationally attempt to make money in the markets). But the long term upward drift of indices is surely not much more than a mixture of inflation and survivor bias (the bust and underperforming companies get weeded out of the indices). More interesting question is what short term features of the market are sufficiently non-random to be worth pursuing.....

And you can still go bust tossing a fair coin, never mind playing the markets.

 

[trendie]

Warning - silly post !!

bonsai - if you believe that markets are quantum related, then surely the next tick is BOTH up and down !!

However, it only becomes one state the moment you place your trade.
Then sure enough, it goes the other way !!

trendie

 

[xetrastock]

If you assume that the stock price is a martingale (a fair game), of which the coin toss is an example, then it is possible to prove that no investment strategy subject to (i) finite wealth and (ii) previsibility (can't see into the future) can bring you expected returns greater than 0. So if you add in trading costs... you'd be very lucky to break even in the long run.

So it seems unlikely, at least to me, that share prices operate in this fashion. I would still say that share prices are random - you can never say that a share price will rise or fall with complete certainty. However, all that is required is a slight edge: If you win your stake with probability 0.5 + e and lose it with probability 0.5 -e, for 0<e<0.5 then you're onto a winner if you play carefully. I guess the question is.... How do you get that edge?

 

[trendie]

Hi xetrastock,

re: coin-flipping edge.

How about a "loaded" coin.

For example, if you believe the market is random, flipping a coin will determine whether you buy or sell, since either option is valid.

However, you load the win/loss.
EG; if the flip says buy;
you buy, you set a stop-loss at 30 pts.
And you also set a stop-win at 33 pts.
( the range would be a function of recent volatility ).

NB: obviously, if the decision is to short, you would set the figures the other way around)

If there is a 50/50 chance of winning or losing, you would win 33pts to every 30pts lost.

Would this work ? Would this give you an edge in the random market ?

trendie

 

[xetrastock]

If you assume that the share price is a fair game (martingale), and you set your stops as above then the probability of the share price hitting your stop win before your stop loss will be lower (since it's further away) than 0.5, and this will compensate for the fact that you win 33 when you're right and lose only 30 when you're wrong. So you're in a break even situation again in the long run. If the share price is a fair game, it really doesn't matter what strategy you employ - you just can't win in the long run! It's lucky that share prices are not martingales!

Somehow you have to gain some insight into where the market will move, with your insight leading to trades that are on average profitable.

 

[trendie]

Sort of tried to consider that by suggesting just that slight 3pt edge.
It would have been silly for me to suggest a 30pt stop-loss for a 60pt stop-win.
( thats obviously silly ).

But yes, what is most likely to happen is I win 30 times ( 30 x 33 ), but lose slightly more often,
( 33 times x 30 points ), resulting in breakeven.
In reality a loss, as the SB spread is part of the stop-win/loss calculation.

I guess I wont be featuring in the next Market Wizards !!

Have a good evening, and look forward to your future posts.

 

[xetrastock]

Yes, I guess this is what makes trading so appealing - it's a real achievement to trade profitably in the long run.

Take care

xetrastock

 

[bonsai]

or if you trade nikkei a fall of over 550pts !


LOL

 

[growltiger]

Yes, I guess this is what makes trading so appealing - it's a real achievement to trade profitably in the long run.

Take care

xetrastock

...also, the short run

 

[SOCRATES]

All this is very entertaining, very funny indeed !

 

[bracke]

Entertaining, funny but is it enlightening?
Are we any further forward or have we simply been engaged in a navel gazing philosophical discussion.
What have we acheived, where is the revised/new trading strategy based on the thoughts put forward?

Regards

bracke

 

[damianoakley]

I've been reading this thread with interest and thought i'd put in a comment or two.

I've actually posted these comments before on other threads, but they are worth repeating.

You can actually make money in the markets by letting a coin flip decide your entry (long or short). Van K. Tharp did numerous tests on this and showed a profit after every test. I've done my own tests and come up with similar results. The thing is, very few people would trade a system based on random entry because it goes against our basic instincs of wanting control over the markets.

The most simple things work best in the markets, and you can't get much simpler than a random entry based on a coin-flip. But most people would not even entertain the idea.

Just my thoughts.

 

[bracke]

damianoakley

What time frame are you talking about?
Does it matter if you sell or buy?

If I flip a coin today and decide heads is buy and tails is sell and I take a long view and am not concerned with the amount of drawdown eventually I will win . It may take some time but the market moves enough to give me a result but is it a practical trading strategy?

Regards

bracke

 

[damianoakley]

bracke

First of all - no particular time frame. The original tests were done on daily charts, but i've tested random entry from daily charts to 5 minute charts.

Does it matter if you sell or buy? - not sure what you mean by this. You flip a coin. If it's heads you go long (buy), if it's tails you go short (sell).

The coin-flip entry does NOT mean taking the long-view and waiting until you show a profit as you suggest. That's not a trading system, it's guesswork.

Once the direction of the trade had been determined by the coin-flip, the market was entered and a definite trailing stop set immediately. If the stop was hit, the coin was flipped again and the market re-entered accordingly.

 

[bonsai]

is there pork in the tree tops ?

 

[damianoakley]

Everyone trades their own beliefs about the markets. What works for one person does not necessarily work for another.

I did say in my post above that most people won't even entertain the idea.

Random entry has been proven to work. If people didn't instantly dismiss the idea and looked into it properly, they too would see that it IS possible to profit from random entry. Remember that it's only the DIRECTION of the trade that is determined by the coin flip. You still control how much is risked and when to exit.

Most people won't look at it at all however because it goes against human nature. They would rather laugh at people like me who suggest the idea!

 

[bracke]

damianoakley

If random entry has been proven to work do you use it yourself?
If yes with what result , if no why not?

Regards

bracke