Does the market need forecasting with more than 50% probability ?
The issue of forecasting has been raised many times, but it is obviously not a topic of much interest to the bulk of traders.
Not everyone can do forecasting, but there are those who can do it. And with the right approach and organization it may even be programmed. But is it really necessary in real life?
For my own statistics in the last two years I have worked out 88.2134% of forecasts in trading. I know about ten people who have had much higher percentage for at least 5 years.
So I'm wondering how to unite these people to work together for further development.
In many threads you see the statement that in order to work in the market the probability of correct prediction must be, well, necessarily greater than 0.5 or >50%. I've always said in such cases that this is a myth, and cited the example of poker, where the probability of winning is ~1/9 -1/6 and good players, however, are always in the black. And I was always told, well... poker is different.
And here is a chance to dispel these well-established legends and myths of our town.)) There is confirmation that in these proverbial 50% absolutely no need.
I am currently working on a new strategy, the first tests have been passed. I haven't performed any optimization, parameter fitting or anything else - it works straight from the page, with initial parameters. No machine learning, neural networks etc. used in strategy. Whether I'll bring it to real - I won't: I have no idea at all, but test results are very interesting.
First of all, for this post I downloaded new data from the Internet and ran a test on it. Secondly, the strategy already took into account all spreads, slippages and so on. On real account it will show better results than in the test.
So, the test results:
Longs:
Deals -407, Profitable - 186, unprofitable - 221, Profit/loss ratio - 0.4570025.
Total profit in longs - 5649, total loss - 2223, profit/loss ratio - 2.5411606.
Shorts:
Deals - 419. , Profitable -182, unprofitable - 237, Profit/loss - 0.4343675,
Total profit in the shorts - 4938, total loss - 2419, profit/loss - 2.0413394
Total longs - shorts:
Deals - 826, Profitable - 368. , Losses - 458, Profit/loss ratio - 0.4455206 ,
Total profit - 10291, total loss - 4642, profit/loss ratio - 2.2169324.
Now for everyone's favorite profit chart. Trading is performed with a fixed lot, and the profit is simply shown on the chart in points, without relation to the trade volume.
At X - trade number, at Y - accumulated profit in pips.
I can provide a CSV file with results of all deals to those who would like to recalculate and check them by themselves. Information about the traded instrument will naturally be deleted). Do not be sorry).
As we can see, in this test, there is absolutely no need in the probability of a correct entry more than 0.5. 0.44 is quite enough, and even more than that.))
Where did you get the idea that profit/loss depends on the ratio of trades? For example, you may have only one profit/loss trade per 100 pips, while you have 9 losing trades of 10 pips and you are in the pole position in spite of the ratio of 1/9. Another thing is if you close equally, e.g. on stops, if you have such a strategy, then show CSV.
The issue of forecasting has been raised many times, but it is obviously not a topic of much interest to the bulk of traders.
Not everyone can do forecasting, but there are those who can do it. And with the right approach and organization it may even be programmed. But is it really necessary in real life?
For my own statistics in the last two years I have worked out 88.2134% of forecasts in trading. I know about ten people who have had this percentage much higher for at least 5 years.
So I'm wondering how to unite these people to work together for further development.
In general it's easy to trade at flat, I'm fooling around with my hands since Feb 21, the initial deposit is $248, the profit is $243. Dumbest work on the channel. What a prediction...
In many threads you see the statement that in order to work in the market the probability of correct prediction must be, well, necessarily greater than 0.5 or >50%. I've always said in such cases that this is a myth, and cited the example of poker, where the probability of winning is ~1/9 -1/6 and good players, however, are always in the black. And I was always told, well... poker's different.
And here is an opportunity to dispel these well-established legends and myths of our town.)) There is confirmation that in these proverbial 50% absolutely no need.
I am currently working on a new strategy, the first tests have been passed. I haven't performed any optimization, parameter fitting or anything else - it's working straight from the page, with initial parameters. No machine learning, neural networks etc. used in strategy. Whether I'll bring it to real - I won't: I have no idea at all, but test results are very interesting.
First of all, for this post I downloaded new data from the Internet and ran a test on it. Secondly, the strategy already took into account all spreads, slippages and so on. On real account it will show better results than in the test.
So, the test results:
Longs:
Deals -407, Profitable - 186, unprofitable - 221, Profit/loss ratio - 0.4570025.
Total profit in longs - 5649, total loss - 2223, profit/loss ratio - 2.5411606.
Shorts:
Deals - 419. , Profitable -182, unprofitable - 237, Profit/loss - 0.4343675,
Total profit in the shorts - 4938, total loss - 2419, profit/loss - 2.0413394
Total longs - shorts:
Deals - 826, Profitable - 368. , Losses - 458, Profit/loss ratio - 0.4455206 ,
Total profit - 10291, total loss - 4642, profit/loss ratio - 2.2169324.
Now for everyone's favorite profit chart. Trading is performed with a fixed lot, and the profit is simply shown on the chart in points, without relation to the trade volume.
At X - trade number, at Y - accumulated profit in pips.
I can provide a CSV file with results of all deals to those who would like to recalculate and check them by themselves. Information about the traded instrument will naturally be deleted). Do not be sorry).
As we can see, in this test, there is absolutely no need in the probability of a correct entry more than 0.5. 0.44 is quite enough, and even with an excess.)
You're getting confused about the probability calculations. You have to consider the ratio of profit to loss.
So you don't have to do anything and cash falls into your pocket? I'm an idiot - I'm studying formulas... That's a shame, gentlemen!
The market is not ruled by formulas but by money, and whoever has more of it is right.)
It's good that you realize your mistakes, it's the key to success!
I'm having cognitive dissonance... Yuri, at least something of higher mathematics to justify the miracle - studios!
In poker, the probability of winning is only 1/9 to 1/6. Learn how to play poker. You should at least start here - world poker club. You'll understand a lot.)
- 2009.10.15
- Разработчик Crazy Panda
- games.mail.ru
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
In many threads you see the statement that in order to work in the market the probability of correct prediction must be, well, necessarily greater than 0.5 or >50%. I've always said in such cases that this is a myth, and gave the example of poker, where the probability of winning is ~1/9 -1/6 and good players, however, are always in the black. And I was always told, well... poker's different.
And here is an opportunity to dispel these well-established legends and myths of our town.)) There is confirmation that in these proverbial 50% absolutely no need.
I am currently working on a new strategy, the first tests have been passed. I haven't performed any optimization, parameter fitting or anything else - it's working straight from the page, with initial parameters. No machine learning, neural networks etc. used in strategy. Whether I'll bring it to real - I won't: I have no idea at all, but test results are very interesting.
First of all, for this post I downloaded new data from the Internet and ran a test on it. Secondly, the strategy already took into account all spreads, slippages and so on. On real account it will show better results than in the test.
So, the test results:
Longs:
Deals -407, Profitable - 186, unprofitable - 221, Profit/loss ratio - 0.4570025.
Total profit in longs - 5649, total loss - 2223, profit/loss ratio - 2.5411606.
Shorts:
Deals - 419. , Profitable -182, unprofitable - 237, Profit/loss - 0.4343675,
Total profit in the shorts - 4938, total loss - 2419, profit/loss - 2.0413394
Total longs - shorts:
Deals - 826, Profitable - 368. , Losses - 458, Profit/loss ratio - 0.4455206 ,
Total profit - 10291, total loss - 4642, profit/loss ratio - 2.2169324.
Now for everyone's favorite profit chart. Trading is performed with a fixed lot, and the profit is simply shown on the chart in points, without relation to the trade volume.
At X - trade number, at Y - accumulated profit in pips.
I can provide a CSV file with results of all trades if you want to recalculate and check them by yourself. Information about the traded instrument will naturally be deleted). Do not be sorry).
As we can see, in this test, there is absolutely no need in the probability of a correct entry more than 0.5. 0.44 is quite enough, and even with an excess.)