Why do I see "Trends" in Randomly Generated Data?

[Corey]

Well, I have always fundamentally justified the existence of trends as the imbalance between buyers and sellers. In this sense, trends do exist. Oil went up because there were more buyers than sellers. Financials went down because there were more sellers than buyers. Sometimes these imbalances are due to news, other times it is just due to emotion.

So while the market may look like a random walk, the trends can be justified. The noise, then, can be justified as the tug-of-war between buyers and sellers.

That does NOT mean, however, that we can define the persistence of trends. Just because a trend exists one minute does not mean there is a higher probability for it to exist in the next. The best thing we can do is enter the trade, jump on the trend, and place stops in areas that identify when the trend has broken (aka there is no longer an imbalance). Stops are fixed distances (be it % or ATRs) say nothing about the current market dynamic. You should be placing your stops in areas that define when the market condition has changed. Often, trends over-extend, which result in a reverse trend, as the imbalance swings wildly from side to side.

So while the market may look like a random walk, I believe that the trends can be fundamentally justified and traded.

But hey, that is just one guys opinion.

 

[PitchBlack]

Quote from Indrionas:

All I was trying to show is that people can see "trends" in the data where there are none. There are no real trends in randomly generated data. This is only to answer to the thread "Why do I see trends in randomly generated data". If you can see "trends" in random data, then how do you know you're not fooling yourself when you're looking at, say, intraday chart.

I know that i'm not fooling myself when i see my bank account and my brokers statements. I daytrade since 1990 and i'm still trading. So i think that the hardproof that i have is not a result of luck; there has to be more. There has to be a system. And a system can only work if things are not completely random.
I believe in behavioural finance. All trades are driving by humans and therefore not random.

But i agree that in real random data you cannot see trends.

 

[Jerry030]

Quote from Corey:

Well, I have always fundamentally justified the existence of trends as the imbalance between buyers and sellers. In this sense, trends do exist. Oil went up because there were more buyers than sellers. Financials went down because there were more sellers than buyers. Sometimes these imbalances are due to news, other times it is just due to emotion.

So while the market may look like a random walk, the trends can be justified. The noise, then, can be justified as the tug-of-war between buyers and sellers.

That does NOT mean, however, that we can define the persistence of trends. Just because a trend exists one minute does not mean there is a higher probability for it to exist in the next. The best thing we can do is enter the trade, jump on the trend, and place stops in areas that identify when the trend has broken (aka there is no longer an imbalance). Stops are fixed distances (be it % or ATRs) say nothing about the current market dynamic. You should be placing your stops in areas that define when the market condition has changed. Often, trends over-extend, which result in a reverse trend, as the imbalance swings wildly from side to side.

So while the market may look like a random walk, I believe that the trends can be fundamentally justified and traded.

But hey, that is just one guys opinion.

There are a lot of other opinions to support your perspective:

From - http://www.economymodels.com/factalmarkets.asp

"The other possibility to explain the fractal dimension is to give up the random walk assumptions. Under a random walk, all changes are independent. At every point in time, the market is equally likely to move up as to move down, regardless of how it has moved before. A fractal dimension below 1.5 could be explained if the market where prone to trending (a straight line, the ultimate trend, would have a fractal dimension of one). A trend is not a random walk, as prices are more likely to move with the trend than against. But the trending would have to be fractal – there would have to be trends at every time scale, and those trends could potentially go in different directions."

The article goes into the fractal nature of the market, which takes
things beyond the basis statistics debate, which is useless.

 

[marketsurfer]

Quote from Corey:

Well, I have always fundamentally justified the existence of trends as the imbalance between buyers and sellers. In this sense, trends do exist. Oil went up because there were more buyers than sellers. Financials went down because there were more sellers than buyers. Sometimes these imbalances are due to news, other times it is just due to emotion.

So while the market may look like a random walk, the trends can be justified. The noise, then, can be justified as the tug-of-war between buyers and sellers.

That does NOT mean, however, that we can define the persistence of trends. Just because a trend exists one minute does not mean there is a higher probability for it to exist in the next. The best thing we can do is enter the trade, jump on the trend, and place stops in areas that identify when the trend has broken (aka there is no longer an imbalance). Stops are fixed distances (be it % or ATRs) say nothing about the current market dynamic. You should be placing your stops in areas that define when the market condition has changed. Often, trends over-extend, which result in a reverse trend, as the imbalance swings wildly from side to side.

So while the market may look like a random walk, I believe that the trends can be fundamentally justified and traded.

But hey, that is just one guys opinion.

why not just randomly enter, if your right, you profit. if your wrong, you lose. random entries, taking quick losses and letting profits run--what's the difference in the end?

surf

 

[straywolfe]

Forgive me if this has already been stated.

Random Number generators are not random. It is impossible for a computer to "flip a coin". A random number equation is designed to produce a sequence of numbers that statistically appear to be random.

The quality of a specific algorithm for random number generation is in large part based on a desired application.

Therefore, in "random" created data, trends and patterns may in fact form over time as the "bigger picture" of the algorithm starts to show it's face.

Having said that, I still find a great deal of the posts in this thread interesting.

 

[PitchBlack]

Quote from Corey:

That does NOT mean, however, that we can define the persistence of trends. Just because a trend exists one minute does not mean there is a higher probability for it to exist in the next.

I don't agree.
I trade a multiple timeframe system and i see, most of the time, the probability of prices moving higher (or lower) for the next minute and even for longer periods. I already demonstrated this to certain ET members in privat ( in privat because i do not like the bashers on ET). This does not mean that i can see the tops and the bottoms (although i can see them in hindsight ), but that's not necessary for profitable trading.

 

[Indrionas]

Quote from spike500:

I know that i'm not fooling myself when i see my bank account and my brokers statements. I daytrade since 1990 and i'm still trading. So i think that the hardproof that i have is not a result of luck; there has to be more. There has to be a system. And a system can only work if things are not completely random.
I believe in behavioural finance. All trades are driving by humans and therefore not random.

But i agree that in real random data you cannot see trends.

I don't doubt or question your success or anything like that.

The whole point was that people tend to see what isn't there. Just google around and you will find lots of TA market "analysis"

I'm not denying that there are some dependencies and non-randomness in the market data at some points in time.

 

[heypa]

I was going to stay out of this esoteric enlightened discussion but have been drawn in. No matter how deeply you look at those combinations of prices that appear to be trends it would be hard to deny that prices in the past have changed. If someone sees a trend leading up to the last day It doesn't make any difference whether it is the result of random or price pressure action
No one can predict the future but you can observe and extrapolate what might happen. I wonder what percentage of you really believe that any obvious trend is more likely to end today than to continue. I think most would expect continuation if forced to put your money where your mouth is.
Do not over analyze. Don't insist on every available bit of information. We humans are prone to believe that the complicated solution to a problem is the better one.It ain't necessarily so. That Franciscan friar from Oakham England observed this many years ago and is remembered.
If you see a trend. Go with. Drop it when it no longer looks like a trend.
It don't matter whether random or non random. It's not how you get in that determines success. It's HOW YOU GET OUT. This ain't rocket science. It's money management.
Take what the market offers. Don't force it.

 

[Corey]

Quote from marketsurfer:

why not just randomly enter, if your right, you profit. if your wrong, you lose. random entries, taking quick losses and letting profits run--what's the difference in the end?

surf

Because if you read anything I wrote, you would see that there is a huge difference. I base entries and exits off of areas that I feel have a high probability of defining a new market characteristic.

For example, we've all seen trend lines fail. Often catastrophically. So I place my stop loss under the trend-line, and place my sell once the trend line has failed a new high has not been achieved. In this situation, I am trading the changing market dynamic.

 

[marketsurfer]

ok. So it cant be defined and is based on feelings?