Random Walk Theory

[BSD]

BSD, dunno what you're getting at in the other thread - a genuinely random walk (with continuous pricing and time) cannot possibly be traded profitably, nor does it trend.

Yes you can, and yes it does.

A random chart looks exactly like any other chart.

Watch this video here that gives a perfect explanation:

Random Charts can be traded profitably Video

The reason why markets are generally random is really very very simple: there is no inner logic to markets, there is no system to markets, there is no secret explanation to price development...

Anything can happen at absolutely any time if a large enough order pushes a market in a direction opposed to what your analysis would have you believe should happen next...

Markets are constantly being pushed to and fro by the diverging interests of all their participants all following their own agenda.

A market is nothing else than a conglomeration of huge numbers of participants all following totally different objectives...

You have hedgers, you have speculators.

You have fundamental traders, you have technical traders.

You have scalpers, daytraders, swing traders, position traders.

You have participants that see the same price levels, yet for some price is too high, for others it's too low.

Etc etc.

Every participant in markets has a different perceptive, different objectives, and different risk parameters.

That is why the notion of predictable markets that follow some inner system is nonsense.

BUT, and that is really the main point that also comes across prefectly in the video above and where conventional wisom got it wrong, markets are generally random, BUT can nevertheless be profitably traded:

Random charts exhibit the exact same properties of charts of real instruments, namely a tendency to develop clear trends.

And it is exactly that that allows profitable trading of BOTH random charts if you want to have some papertrading fun, AND of real charts.

If you have somebody claiming they've come up with a great new system thats gonna beat markets should just watch Gamma Jammer short 1 Billion worth of EUR/USD to see price bias change sufficiently to knock out their friends analysis as no more than the latest futile holy grail search.

One does not need a holy grail.

All one needs is the ability to see where the path of least resistance is, jump on, and hang on until the trend bends, simple cutting ones losers and letting ones winners run kind of stuff.

Anybody who claims they've got a holy grail will

A: never be able to explain what they do which always means they do not understand what they are on about themselves

and

B: never be able to back up their claims with a public track record.

Public track records say one thing which clearly debunks the notion that it is possible to significantly outwit markets:

Kenneth Grant, in "Trading Risk: Enhanced Profitability through Risk Control", depicts his experience as risk manager for some of the best and most successful hedge funds, amongst others Paul Tudor Jones funds and Steve Cohens SAC Capital, that:

"ACROSS ALL TRADING STYLES, TIME FRAMES AND TRADERS, ONE RULE HOLDS TRUE:

10% OF ALL TRADES INEVITABLY ACCOUNT FOR 90% OF PROFITS !"


QED.

 

[arabianights]

A random chart (in the sense of a random walk) does not have trends because the past will not affect the future. You may see things that look like trends, but they are not trends.

It's pretty academic anyway as I'm not aware of any market that follows a random walk.

Also it's very easy to trade/invest if you believe the market follows a random walk. Decide how much return you want, leverage yourself appropriately, and buy indiscrimnately. Wait two years and blow up on a "ten sigma" event. Rinse and repeat with the same shareholders.

 

[black bear]

Classic

Split's Random Walk Theory.

For what it's worth.

Once a trend has been established in a large timeframe--daily, for instance-- the smaller timesframes operate on a random basis.

They can be very profitable but they are random, nevertheless, and operate within longer trends.

This is one of the reasons that fairly close stops must be used because, no matter what the TA pundits say, there is no telling how a price is going to act during a morning session, nor any reason why it should not reverse after lunch, when the Dow opens. In fact, it can reverse several times during the day and, still, close in line with the daily trend.

If anyone disagrees then, please, explain the antics of the SBC's during the first hour of trading, when they are trying to establish a price. Spreads and prices are up the wall, most times. That is because no direction has been established by the traders.

Split

I'm officially opening a Split's Random Walk Theory Fan Club with me as the first member

BSD,

I already started the fan club filed under trend lines and no mans good enougth for 2 :?:

are they :-0

great post Split another classic :smart:


and again for all those that have missed it before

just a bit of a FUSS for you Split :clap:


Trend lines, to me, are drawn arbitrarily. I just use them as a sort of "comfort line". If I am on the right side of it I feel fine and have no worries. The rule that they must be drawn across the tops or bottom is, more or less, telling the market what to do. How can you expect that of a price? When the price cuts the trend I look to see if a pattern is forming. If the pattern is continuous, I revert to a trend line, again, almost certainly, the momentum will have changed but if I keep to the right of it in a bear and to the left of it in a bull then I am OK- Whenever it crosses, it is a warning to pay attention to the trade, nothing more.

There is nothing to be learned from trend lines, IMO anyway, except to help keep the ship on course. Go inside it and you are entering into shallow water.

 

[BSD]

You may see things that look like trends, but they are not trends.

I'm not really harping on about whether markets are random or not.

My main point is simply that it is impossible to predict what happens next, that the only people who claim they can do that could never explain what they do, nor ever back it up with a public track record.

Real life, public track records show what's what: the best of the best show that no one manages to beat a distribution that has 10% of trades generating 90% of profits.

If someone had figured out how to beat markets they could improve that ratio, which just does not exist.

But, and that is the most important point, one does not need a holy grail to make more money from trading than one could ever spend.

People just waste a lot of time and energy erroneously believing that if they just keep investing more sweat and midnight oil on studying markets they'll eventually discover the secret.

They won't, because there is no secret.

Trading is just a numbers game.

I already started the fan club filed under trend lines and no mans good enougth for 2 :?:

are they :-0

 

[BSD]

Public track records say one thing which clearly debunks the notion that it is possible to significantly outwit markets:

Kenneth Grant, in "Trading Risk: Enhanced Profitability through Risk Control", depicts his experience as risk manager for some of the best and most successful hedge funds, amongst others Paul Tudor Jones funds and Steve Cohens SAC Capital, that:

"ACROSS ALL TRADING STYLES, TIME FRAMES AND TRADERS, ONE RULE HOLDS TRUE:

10% OF ALL TRADES INEVITABLY ACCOUNT FOR 90% OF PROFITS !"


QED.

LOL, for the tenth time.

 

[firewalker99]

People just waste a lot of time and energy erroneously believing that if they just keep investing more sweat and midnight oil on studying markets they'll eventually discover the secret.

They won't, because there is no secret.

Are you sure? If you looked at the best selling books (oh wait, there's also the DVD!) of 2007, one might think there is.

 

[BSD]

Are you sure? If you looked at the best selling books (oh wait, there's also the DVD!) of 2007, one might think there is.

LOL ! :cheesy:

 

[firewalker99]

I've explained more than clearly why there cannot be any secret to a market that is composed not of an inner logic or system that would be seperate from it's constituting participants, no, a market is nothing but the sum of it's constituting participants, each of whom has his own agenda, doing his own thing.

There are those who believe that 'external factors' which have an impact on us all influence the decisions we make as individuals. Hence we might think we are making decisions for ourself, but basically we are governed by another force of which only the effects are visible (like gravity). I'd even go so far as saying it comes down to performing an exploratory factor analysis to identify if such variables exists. We are thus looking for yet unidentified unobserved causal factors. Unfortunately on the scale we are talking about, this isn't a very realistic task to perform.

Of the above there seems to be a subgroup, who believe the lunar cycles is one of those external factors. Incidentally, this thread might interest you guys as well:

http://www.trade2win.com/boards/gen...illes-welder.html?highlight=lunar+moon+cycles

 

[Splitlink]

There are those who believe that 'external factors' which have an impact on us all influence the decisions we make as individuals. Hence we might think we are making decisions for ourself, but basically we are governed by another force of which only the effects are visible (like gravity). I'd even go so far as saying it comes down to performing an exploratory factor analysis to identify if such variables exists. We are thus looking for yet unidentified unobserved causal factors. Unfortunately on the scale we are talking about, this isn't a very realistic task to perform.

Of the above there seems to be a subgroup, who believe the lunar cycles is one of those external factors. Incidentally, this thread might interest you guys as well:

http://www.trade2win.com/boards/gen...illes-welder.html?highlight=lunar+moon+cycles

Are these what the Spanish call "lunatic cycles" because, if they are, its quite a big sub-group you're talking about.

 

[black bear]

Splits back

Are these what the Spanish call "lunatic cycles" because, if they are, its quite a big sub-group you're talking about.

Hi Split

your on form today

Is this the group we make our living off Split :?:

have a good week if not already, off out in a moment as per usual, but ftse is holding my attention at the minute....................again

 

[FXSCALPER2]

A chart generated from random phenomena doesn't itself has to be random. A true randomwalk like a toss of a coin will produce a chart that will look like a stock chart because the principles gooverning the two are similar. This is, by the way, knowable a priori. A truly random event is serial and you have to look at each event to see the randomness. If you bunch the events in larger chunks, which is what charts do, then due to large diviations which take time to disappear, you will get recginisable and tradable patterns. This is the reason why trading is not gambling: the data are not (under normal circumstances) rigged.

It is misleading to look at random charts and compare it to a random walk chart like a coin toss because you really do not trade the random walk itself but the chart of the random walk which needn't be itself random.

 

[firewalker99]

A chart generated from random phenomena doesn't itself has to be random. A true randomwalk like a toss of a coin will produce a chart that will look like a stock chart because the principles gooverning the two are similar. This is, by the way, knowable a priori. A truly random event is serial and you have to look at each event to see the randomness. If you bunch the events in larger chunks, which is what charts do, then due to large diviations which take time to disappear, you will get recginisable and tradable patterns. This is the reason why trading is not gambling: the data are not (under normal circumstances) rigged.

It is misleading to look at random charts and compare it to a random walk chart like a coin toss because you really do not trade the random walk itself but the chart of the random walk which needn't be itself random.

So are you saying that although coin tosses themselves are random, the chart of a coin toss experiment is not?

Have a look at the attached XLS file (you need to enable macro's for this to work). You can generate a chart based on flip coins.

 

[arabianights]

It is misleading to look at random charts and compare it to a random walk chart like a coin toss because you really do not trade the random walk itself but the chart of the random walk which needn't be itself random.

But if the chart of the random walk was produced randomly, then you cannot trade it because the future walk cannot be predicted from its previous path. The previous path has no predictive value. You have an expectancy of zero. That's the whole point!

 

[dbphoenix]

Perhaps if you guys were to explore the difference between randomness and unpredictability? :whistling

 

[FXSCALPER2]

People are making what philosophers call a 'category error'. It is said that the average person has 2.5 children. No one actually has 2.5 children, of course. Similarly, although a coin toss is a simple random walk, that randomness does disappear depending on what you are looking at. Your expectancy is zero only if you bet on each throw. What you can do is bet on frequency. If you get 10 heads in a row, for example, and each head means one pip up, it will take take a long time before the trend changes simply because it is more unlikely than likely in any time window for there to be ten tails so you can bet on an up trend using the beginning of the 10 head series as support. Incidentally, that is the reason why you see a trend in a randoom series. The fascinating thing is, if you actually look at the outcomes of each throw in the series that represents a trending chart of a coin toss experiment, it is highly likely that the frequency was close to 50:50. That looks paradoxical only if you do not separate in your head the individual outcomes from what you see on the chart.

I am sure if you look at what actually happened in the day to day movement of the entire uptrend of the last seven years in EURUSD, the indvidual outcomes are much more evenly balanced than the monthly chart suggests.If you look tick by tick, you will probably end up with close to 50:50 up and down. The reason why you see a trend is because of clusters and they do happen in a random walk.

 

[arabianights]

People are making what philosophers call a 'category error'. It is said that the average person has 2.5 children. No one actually has 2.5 children, of course. Similarly, although a coin toss is a simple random walk, that randomness does disappear depending on what you are looking at. Your expectancy is zero only if you bet on each throw. What you can do is bet on frequency. If you get 10 heads in a row, for example, and each head means one pip up, it will take take a long time before the trend changes simply because it is more unlikely than likely in any time window for there to be ten tails so you can bet on an up trend using the beginning of the 10 head series as support. Incidentally, that is the reason why you see a trend in a randoom series. The fascinating thing is, if you actually look at the outcomes of each throw in the series that represents a trending chart of a coin toss experiment, it is highly likely that the frequency was close to 50:50. That looks paradoxical only if you do not separate in your head the individual outcomes from what you see on the chart.

I am sure if you look at what actually happened in the day to day movement of the entire uptrend of the last seven years in EURUSD, the indvidual outcomes are much more evenly balanced than the monthly chart suggests.If you look tick by tick, you will probably end up with close to 50:50 up and down. The reason why you see a trend is because of clusters and they do happen in a random walk.

The further away you put it, the less chance you have of your stop being hit before your position moves into profit. So more of your trades will win. Unfortunately your losses will be bigger, so this doesn't change your expectancy which remains 0.

 

[temptrader]

Randomness is an objective property. Nevertheless, what appears random to one observer may not appear random to another observer who has the key needed to turn the sequence of bits into a readable message.

Sorry, but randomness is NOT a subjective property. There might be issues in not getting enough data to, how shall we put it, say with a 98% confidence interval that the data is random. There are objective measures we can use to put the data to that test.

I suppose in this sense, practically speaking anyway, that you can never be able to prove that something is random because you do not have infinite data for you to put through your objective measures.:cheesy:

 

[dbphoenix]

Sorry, but randomness is NOT a subjective property. There might be issues in not getting enough data to, how shall we put it, say with a 98% confidence interval that the data is random. There are objective measures we can use to put the data to that test.

I suppose in this sense, practically speaking anyway, that you can never be able to prove that something is random because you do not have infinite data for you to put through your objective measures.:cheesy:

What I quoted did not say that randomness is a subjective property; it said the opposite. It then addressed what appears to be random but may in fact not be.

 

[darktone]

who " really " believes this...? :|

Nah!
Its an auction. How can an auction be random!?

 

[temptrader]

What I quoted did not say that randomness is a subjective property; it said the opposite. It then addressed what appears to be random but may in fact not be.

terribly sorry, must have been dyslexic there for a minute.