Most sucessful? - page 5
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
I hesitate because, being I am not a programmer, one of my ea's started to gobble up my disk to 99% usage! I fixed the problem due to the graces of the competent and expert programmers on this site.
What I can do is tell you how to write the ea and what market condition it is monitoring. It is a simple ea and you could code it in one afternoon.
I probably will do this soon when I have time to explain the method.
Used all available date.
It might not be enough. Serious studies can span years.
Since market can change even with an unlimited large sample you cannot be confident
But you must be confident if you make a claim of correlation. There is either a correlation or there is not. If you admit you cannot be confident, then your study needs more data. If you took your "experiment" to a professor of statistics, he or she would demand a more rigorous research. You might be right, but you need more data to demonstrate it.
No trend trader would enter a trade if he thinks that the market is not going in his direction, this implies they are predicting
No, they are hoping, not predicting. There is no implied prediction if you take a trade after a crossover because you cannot predict that this is going to be the big one. It might only last 5 minutes. Being that you cannot predict the extent of the move, you take the trade and wait to see what happens. There is still an uncertainty in taking this trade, not an absolute insight into the future. If it doesn't move the way you want, you exit. A crossover is a potential opportunity, not a sure flag of the start of a trend.
I would not bet on this
Like all things technical indicators can be used or misused. They are used properly if they are viewed only as a summary or reflection of what the market is doing now in the present tense. For example, candlesticks are excellent descriptors of present and past market action, but they are not predictors of future market action. What you know now in the present cannot tell you what will happen next. A good working definition of predictors is that they predict 100% correctly the future price action. If an indicator cannot do this, it is not a predictor. Technical indicators can reveal probabilities. Stochastics can show overbought for days, but not tell us when the market will reverse.
Whats the difference between trading probabilities and trading predictions?
Good question. Here is how I look at it. Suppose you have an evenly balanced coin. If I flipped it one thousand times, using the laws of probability and the law of large numbers, I could say it will show up heads about fifty per cent of the time and tails fifty per cent of the time. This is using probability. If I said it will show up heads on toss 2,4,6,8, or all the even tosses I am using prediction. Probabilities give fuzzy outcomes, predictions give exact events.
since you are striktly against precitions you cannot make the abore precition. Following striktly the math there is the probability that the crossover will never occur. Yeah, a crossover is very likely to happen, but as always when dealing with probabilities not guaranteet .
A crossover of the 20 ma occurs because of deeper marcro-economic and fundamental forces. There are many plausible explanations. One is the simple supply and demand explanation. Forex is just a mechanism to facillitate international trade. Without the existence of international banks dealing in foreign currencies, world trade would come to a grinding halt. As long as China needs oil and cement, South America needs U.S. pharmaceuticals, the Eurzone needs American technology, Japan needs foreign grains, America needs Canadian oil and on and on, there will be a flow of goods and services flowing between nations. This flow itself will move in macro and micro cycles. This will cause shortages and surpluses, higher demands and lower demands, higher supplies and lower supplies untl price discovery and equilibrium occurs in the market. But equilibrium in the world trade market is as transient as trying to balance a pencil on its point. Surpluses and shortages will immediately emerge again. This will cause foreign exchange rates to fluctuate between nations that participate in floating currency rates (not artificially pegged currencies). This fluctuation will reflect in changing moving averages. Moving averages are lagging indicators, and therefore, current prices will always cross above or cross below these averages. This discussion does not even address the capital flow of money across nations. Yes, if global trading stopped, moving average crossovers would stop. That's predicatable.
Whats your benefit? Since you do not know when it can take years or decades.
True. If you trade only on a weekly or monthly crossover, you could wait for a very long time before a crossover occurs. Meanwhile I would trade on daily charts. You might have to wait two or three weeks before a crossover occurs, but the real benefit in all of this is emotional. It forces you to learn patience, and impatience ruins many a potentially successful trader. I admit I struggle with this. I have scalped before because I am a restless type. But I am learning to be patient because this is required for long term trend trading.
If you are assuming that there is no way to predict the market the probabillity is (due spread + commision) ALWAYS against you, which implies a negative expectation value ALWAYS!!!!
That's harsh and apocalyptic. If that were true, liquidity in all markets-- forex, futures, equities-- would dry up. This would slow down foreign trade and put a damper on hedging in the grains market. That would be catastrophic for large-volume producers and suppliers. A high positive expectation value is the only reason speculation works in the long term. Without a high expectation value there would be no incentive to take on the risk of trading. But traders who do not understand the importance of the laws of probability and the expectation value are inevitably sifted out of the process with empty accounts. Traders who do understand can make a nice living. There are trainloads of them in the world, and unfortunately a small minority. I think people armed with the right information can change that.
It might not be enough. Serious studies can span years.
In my case i used a 10 Years Dataset.
No, they are hoping, not predicting. There is no implied prediction if you take a trade after a crossover because you cannot predict that this is going to be the big one. It might only last 5 minutes. Being that you cannot predict the extent of the move, you take the trade and wait to see what happens. There is still an uncertainty in taking this trade, not an absolute insight into the future. If it doesn't move the way you want, you exit. A crossover is a potential opportunity, not a sure flag of the start of a trend.
For me it is the same saying the probability dictated that the bigger part of the crossover signals go in my direction and : oh a crossover happend, i predict a further movement in my direction. Please notice here that when i use predict i do not imply a 100% winning rate. A good prediction must have a winningrate to beat the 'house edge' (spread+commision)
Like all things technical indicators can be used or misused. They are used properly if they are viewed only as a summary or reflection of what the market is doing now in the present tense. For example, candlesticks are excellent descriptors of present and past market action, but they are not predictors of future market action. What you know now in the present cannot tell you what will happen next. A good working definition of predictors is that they predict 100% correctly the future price action. If an indicator cannot do this, it is not a predictor. Technical indicators can reveal probabilities. Stochastics can show overbought for days, but not tell us when the market will reverse.
Again our disagreement is that we define prediction different. Might be my fault if use this term in the wrong sense.
Good question. Here is how I look at it. Suppose you have an evenly balanced coin. If I flipped it one thousand times, using the laws of probability and the law of large numbers, I could say it will show up heads about fifty per cent of the time and tails fifty per cent of the time. This is using probability. If I said it will show up heads on toss 2,4,6,8, or all the even tosses I am using prediction. Probabilities give fuzzy outcomes, predictions give exact events.
Hm, since we are at coinflipping: Lets assume you flipped tails 20 times in a row, (which is already not very likely) and you know that the probability for 21 times tails in a row is very low, would you take the bet when you have to pay 1% commision to the house? Probably no because you know that the coin does not remember the last 20 flippes and the odds of the bet are still 50:50.
A crossover of the 20 ma occurs because of deeper marcro-economic and fundamental forces. There are many plausible explanations. One is the simple supply and demand explanation. Forex is just a mechanism to facillitate international trade. Without the existence of international banks dealing in foreign currencies, world trade would come to a grinding halt. As long as China needs oil and cement, South America needs U.S. pharmaceuticals, the Eurzone needs American technology, Japan needs foreign grains, America needs Canadian oil and on and on, there will be a flow of goods and services flowing between nations. This flow itself will move in macro and micro cycles. This will cause shortages and surpluses, higher demands and lower demands, higher supplies and lower supplies untl price discovery and equilibrium occurs in the market. But equilibrium in the world trade market is as transient as trying to balance a pencil on its point. Surpluses and shortages will immediately emerge again. This will cause foreign exchange rates to fluctuate between nations that participate in floating currency rates (not artificially pegged currencies). This fluctuation will reflect in changing moving averages. Moving averages are lagging indicators, and therefore, current prices will always cross above or cross below these averages. This discussion does not even address the capital flow of money across nations. Yes, if global trading stopped, moving average crossovers would stop. That's predicatable.
This of course is all true
True. If you trade only on a weekly or monthly crossover, you could wait for a very long time before a crossover occurs. Meanwhile I would trade on daily charts. You might have to wait two or three weeks before a crossover occurs, but the real benefit in all of this is emotional. It forces you to learn patience, and impatience ruins many a potentially successful trader. I admit I struggle with this. I have scalped before because I am a restless type. But I am learning to be patient because this is required for long term trend trading.
That is really true, patience is your biggest weapon and impatience your worst enemy, most trades have learnd that first hand ;)
That's harsh and apocalyptic. If that were true, liquidity in all markets-- forex, futures, equities-- would dry up. This would slow down foreign trade and put a damper on hedging in the grains market. That would be catastrophic for large-volume producers and suppliers. A high positive expectation value is the only reason speculation works in the long term. Without a high expectation value there would be no incentive to take on the risk of trading. But traders who do not understand the importance of the laws of probability and the expectation value are inevitably sifted out of the process with empty accounts. Traders who do understand can make a nice living. There are trainloads of them in the world, and unfortunately a small minority. I think people armed with the right information can change that.
Again it seems that we do not agree about the terms. for me expectation value = (winning% * win$) / (loosing% * loose$). The only way to gain in the markets is to have a system which spits out signal with a higher probability to win than to loose.
As conclusion i want to make sure that i really appriciate this argument.
I am sorry if i used the wrong terms sometimes and this leads to misunderstandings.
There are some problematic issues here. To begin with any monitoring of data involves historical data. The moment “current” data is recorded, it becomes past data, and past data can never predict the future price action of the market. Never. No matter how sophisticated the analysis is.
In Statistitic they say 'if the past continues to behave the same into the future then with x% confidence the predicted pattern of behavour will happen p% of the time.
If one could show a deterministic relation between past price action and future price action, that person would receive the Nobel prize in economics and become a very rich person. As sucessful traders already do so.
We should learn some lessons from successful trend traders. They never try to predict the future price action of the market, they just follow the market. They ignore all the future fables that come out of Wall Street and the sensational-appeal-to-the-emotions media. Following the trend is predicting future price action, that is that the trend will remain the same until its not, lets look for a bat or a gartley or a head n shoulders.
Silliness in Wall Street, for example, is ubiquitous. Some Wall Street firm or columnist or guru might say the S&P 500 is going to hit this level in April and it does hit that level. And before you know it people are exclaiming the genius of that firm or person.
This is just as silly as saying, “Flip that coin and I predict it will land on heads”, and what do you know? It lands on heads! And everyone will say, “You must be a genius!” Who said predictions had to be right all the time if you can pick a 20 to 1 horse 18 times out of 20 you are making money.
These people don’t understand the laws of probability and the law of large numbers—that there can be a large variance from the probable outcome if your sample is too small. That firm or person would have to make about 100 correct predictions to be rightfully declared as a genius. It seems to me that you reflecting on the nature of randomness here not probability. The student-t test is used to test hypothesis on small sample sizes all the time.
Now there is a real problem if an ea is monitoring the “current” conditions of the market to determine what the trading activity of that ea should be. If the ea is monitoring current conditions it will be taking too small of a sample because “current” by definition limits the time domain to a small increment. Current is what ever you need it to be to make a decision
What does current mean anyway? One minute? One hour? One day? But is a day a large enough sample? Probably not, so the ea uses a larger time frame. It takes a broader time sample of the market to get more reliable data. But when it does this, the ea is dealing with historical data, and notwithstanding it is a sufficiently large sample, this past data cannot predict the future action of the market. This is the inescapable dilemma: Current means too small of a sample, historical means data which is unreliable and non-correlated with the future. If some sort of consistant prediction is impossible there would not be sucessful traders.
As conclusion i want to make sure that i really appriciate this argument.
I am sorry if i used the wrong terms sometimes and this leads to misunderstandings.
In my case i used a 10 Years Dataset.
Ten years is a good dataset. Maybe I don't totally understand what you did. Please explain in layman terms what your study demonstrated.
Probably no because you know that the coin does not remember the last 20 flippes and the odds of the bet are still 50:50.
Correct. If I believed otherwise I would be committing the gambler's fallacy.
Again our disagreement is that we define prediction different. Might be my fault if use this term in the wrong sense.
Much debate often is due to the ambiguities of language. We humans are in the prediction business all the time even if we don't know it. Every time we make an assumption, we are predicting. And our lives are full of assumptions all the time. When my wife hands me a shopping list and I go grocery shopping for my family, I am assuming my car will start and I will drive safely to and back from the store. I am making these assumptions at a sub-conscious level. I am predicting events that have not yet happened even though I may not be conscious of it.
As traders we have to be careful of this, especially assumptions we make at the sub-concious level. A while a go a friend of mine said he was investing in gold. I asked why and he said because gold will always go up. And I said, but what if it doesn't? He just said, it will always go up. I didn't say anthing else such as, you're trying to predict the future because I know that would have irritated him.
Well, a few weeks later gold took a pretty good dive and I could tell he was upset about it. (He was a victim of the dot.bomb crash years ago and lost almost half of his money.) I was surprised that he was upset. Didn't he learn from his past experiences? Didn't he learn anything I taught him about trading? I taught him previously, don't invest, trade. Investors buy and hold because they are predicting the price will always go up. Traders follow the market long or short because they don't predict it will always go up. But he was following a deeply rooted prediction at the subconscious level that gold would continue to go up forever. He and others should learn the history of speculative bubbles. The Holland Tulip bubble is a case in point. The U.S. housing market is another. All traders should read Michael Covel's book, Trend Trading. It covers all of this very well.
I think many inexperienced traders experience shell shock when their trades go against them because they believe at the sub-conscious level that indicators can predict future price action. They feel cheated when the indicators don't work. It isn't their indicators that have cheated them, it's their own thinking.
I believe the moving average crossover is not a technical indicator. It is a direct price summary signal given by the market that the market is changing its trend. The new trend might only last 5 minutes, but even a scalper in theory can benefit from this. Of course, no one knows if the trend will continue more than 5 minutes. (As I said previously, using a practice account, I made almost a 50% roi in about 20 days using a small 20 ma crossover event. Of course, I would not have done this with a real account because the method I used was extremely risky.)
Again it seems that we do not agree about the terms. for me expectation value = (winning% * win$) / (loosing% * loose$). The only way to gain in the markets is to have a system which spits out signal with a higher probability to win than to loose.
I think you mean: expectation value = (winning% * win$) - (loosing% * loose$). I agree with you if that's what you mean. This formula should be in chapter 1 of every book on forex treading, but it is not. Might be the reason why 95% of all forex traders lose their money. If you haven't seen my spread sheet it's all based on the expectation value formula. See attached spreadsheet. Use green squares only to enter data.
In Statistitic they say 'if the past continues to behave the same into the future then with x% confidence the predicted pattern of behavour will happen p% of the time.
Precisely. No on knows if it will or if it will not continue. "If" admits you cannot know the future. No one can know the future except Merlin, whose time sequence was reversed, who rembered the future and predicted the past. But then, Merlin is a fictional character.
It seems to me that you reflecting on the nature of randomness here not probability. The student-t test is used to test hypothesis on small sample sizes all the time.
I find it hard to believe that any statistical study with a small sample has any significance. Come on, do you believe that? I want to know the mean weight of boulders in the Grand Canyon and the mean weight of boulders near the top of Mt. Rainier. I weigh 3 boulders in the Grand Canyon and get an average and then weigh 3 boulders from Mt. Rainier and get an average there, and then I declare I know boulders are larger in the Grand Canyon than on Mt. Rainier. I don't think so. A small sample tells us nothing but random information.
Who said predictions had to be right all the time if you can pick a 20 to 1 horse 18 times out of 20 you are making money.
I think prediction implies exactness. If you admit you don't always know the exact outcome, then you are using probabilities. Probabilities work for you if you have a good expectation value.
Current is what ever you need it to be to make a decision
You are a general. Your scouts report to you that they saw two enemy soldiers moving toward your left flank. This could mean, the scout tells you, that the enemy is trying to outflank you. You ask, how many more did you see in the past hour? The scout says, only the two. Do you order a full counter-measure of 10,000 infantry on that piece of information the scout gave you? Remember the wisdom of Sun-Tzu: "All warfare is based on some form of deception. If you are weak, make the enemy think you are strong. If you are strong, make the enemy think you are weak." Could we be deceived by the market if we act too soon on a small amount of data?
Precisely. No on knows if it will or if it will not continue. "If" admits you cannot know the future. No one can know the future except Merlin, whose time sequence was reversed, who rembered the future and predicted the past. But then, Merlin is a fictional character.
What reason do you have for thinking that future behaviour is going to be any different from past behaviour?
I find it hard to believe that any statistical study with a small sample has any significance. Come on, do you believe that? I want to know the mean weight of boulders in the Grand Canyon and the mean weight of boulders near the top of Mt. Rainier. I weigh 3 boulders in the Grand Canyon and get an average and then weigh 3 boulders from Mt. Rainier and get an average there, and then I declare I know boulders are larger in the Grand Canyon than on Mt. Rainier. I don't think so. A small sample tells us nothing but random information.
I am sure manufactures would love to wast time with large numbers prooving that there is something wrong or would they be happier to dectect a problem in the shortest space of time possible!
I think prediction implies exactness. If you admit you don't always know the exact outcome, then you are using probabilities. Probabilities work for you if you have a good expectation value.
Here in lies the problem of how different people decide that they are going to use language and demand others use it the same way they do when clearly other groups use language differently and undersand differently.
-------------------------------------------------
Current is what ever you need it to be to make a decision
You are a general. Your scouts report to you that they saw two enemy soldiers moving toward your left flank. This could mean, the scout tells you, that the enemy is trying to outflank you. You ask, how many more did you see in the past hour? The scout says, only the two. Do you order a full counter-measure of 10,000 infantry on that piece of information the scout gave you? Remember the wisdom of Sun-Tzu: "All warfare is based on some form of deception. If you are weak, make the enemy think you are strong. If you are strong, make the enemy think you are weak." Could we be deceived by the market if we act too soon on a small amount of data?
The above example doesn't change or refute the idea. People allow themselves the opportunity to change for you hear the pharase 'I currently believe' meaning the data they have now has brought about their current belief.
What reason do you have for thinking that future behaviour is going to be any different from past behaviour?
Observation and understanding. Sit at a beach and look at the waves crashing in. What is the same? The wave patterns. What is different? The subtle variations of each wave never being exactly the same as the previous. So, the waves are the same in form, but they are different in the details. The details in the market are never the same, only the form. But it is the details that can make or break your account in a trade. You can't predict the details, only the form. I have stated previously that if a price is below a 20 period moving average, it will eventually end up above the average. This is a prediction about a market pattern which is due to the fundamental forces that drive world trade, but not a prediction about the details. No one knows when this crossover will occur and to what extent.
I am sure manufactures would love to wast time with large numbers prooving that there is something wrong or would they be happier to dectect a problem in the shortest space of time possible!
A manufacturer has to always keep an eye on the bottom line to survive in a free market competitive environment. And it would be expensive to have a rigorous quality control system in place with extremely low failure rates as a goal on their outgoing products. But suppose the manufacturer makes a drug that can save the lives of millions of people, but only if each capsule going out of the plant is pure and uncontaminated. Would it be a waste of time to shoot for a 0% defect rate? What if a capsule were contaminated and it could potentially kill a user? Clearly, there would be no room for shortcuts here. Tiny samplings would be unacceptable.
Here in lies the problem of how different people decide that they are going to use language and demand others use it the same way they do when clearly other groups use language differently and undersand differently.
That's why we need to defer to a universal language, such as mathematics, where everyone understands the meanings. The expectation value formula is a perfect example of something with unambiguous terms and meaning. When we talk about trading we should talk about it in terms of the expectation value. Uncertainty is implied in the formula, but success is possible if an ea following a trading method produces a win rate times the gain greater than the loss rate times the loss.
Again it seems that we do not agree about the terms. for me expectation value = (winning% * win$) / (loosing% * loose$). The only way to gain in the markets is to have a system which spits out signal with a higher probability to win than to loose.
I think you mean: expectation value = (winning% * win$) - (loosing% * loose$). I agree with you if that's what you mean. This formula should be in chapter 1 of every book on forex treading, but it is not. Might be the reason why 95% of all forex traders lose their money. If you haven't seen my spread sheet it's all based on the expectation value formula. See attached spreadsheet. Use green squares only to enter data.
Expectation value as you call it would only be positive if (winning% * win$) >(loosing% * loose$) and by normal math manipulation (winning% * win$) / (loosing% * loose$)>1 and perhaps more interesting is the result
(winning%*win$)/loosing%>loose$
What reason do you have for thinking that future behaviour is going to be any different from past behaviour?
Observation and understanding. Sit at a beach and look at the waves crashing in. What is the same? The wave patterns. What is different? The subtle variations of each wave never being exactly the same as the previous. So, the waves are the same in form, but they are different in the details. The details in the market are never the same, only the form. But it is the details that can make or break your account in a trade. You can't predict the details, only the form. I have stated previously that if a price is below a 20 period moving average, it will eventually end up above the average. This is a prediction about a market pattern which is due to the fundamental forces that drive world trade, but not a prediction about the details. No one knows when this crossover will occur and to what extent.
I am sure manufactures would love to wast time with large numbers prooving that there is something wrong or would they be happier to dectect a problem in the shortest space of time possible!
A manufacturer has to always keep an eye on the bottom line to survive in a free market competitive environment. And it would be expensive to have a rigorous quality control system in place with extremely low failure rates as a goal on their outgoing products. But suppose the manufacturer makes a drug that can save the lives of millions of people, but only if each capsule going out of the plant is pure and uncontaminated. Would it be a waste of time to shoot for a 0% defect rate? What if a capsule were contaminated and it could potentially kill a user? Clearly, there would be no room for shortcuts here. Tiny samplings would be unacceptable.
Here in lies the problem of how different people decide that they are going to use language and demand others use it the same way they do when clearly other groups use language differently and undersand differently.
That's why we need to defer to a universal language, such as mathematics, where everyone understands the meanings. The expectation value formula is a perfect example of something with unambiguous terms and meaning. When we talk about trading we should talk about it in terms of the expectation value. Uncertainty is implied in the formula, but success is possible if an ea following a trading method produces a win rate times the gain greater than the loss rate times the loss.
At this point I can no longer be bothered to discuss your fatuous and tangental examples to what I already know happens in the real world.