Hateful pipsqueak. - page 4
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
..... Different sets of take and stops for "hammer" lose on average, as there is no stat advantage, even though some sets are winning for the period.....
Didn't you try it without takeaways and stops ?
Tests have shown that pips with a target of less than 10 pips on EURBucks is not justified. Although it is still positive.
But 10-20 is going very well.
===== Deleted=====
I thought and thought and changed my mind. :(
Wo - I remembered another "anti-science slogan" (like "pseudoscience" - not mine) - "the criterion of truth is social and historical practice". :)
Maybe I don't understand what I should see - tell me, because, unfortunately, the article confirms my skepticism - lack of statistical advantage of the popular "hammer" candle[...] But for unstable minds the article is harmful, because it creates an illusion that "hammer" is a reversal candle according to statistical analysis, not a fitting one.
Yes, Vita, you are quite adequate in your assessment of the article(sergeev, hang in there!): some conclusions are indeed hasty and based on too small a sample size. But the article has an honest attempt to create a system for evaluating indicators.
OK, a counter question: what do you think the methodology for testing of these indicators should be to convince you? What should be the sample size (the number of the cases of the condition fulfillment) for which the ratio of losses vs. profits, say, 40/60 would convince you that there is a statistical advantage?
===== Deleted=====
I thought and thought and changed my mind. :(
Wo - I remembered another "anti-science slogan" (like "pseudoscience" - not mine) - "the criterion of truth is social and historical practice". :)
Yes, Vita, you are quite adequate in your evaluation of the article(sergeev, hang in there!): some conclusions are indeed hasty and based on too small a sample size. But the article has an honest attempt to create a system for evaluating indicators.
OK, a counter question: what do you think the methodology for testing of these indicators should be to convince you? What should be the sample size (the number of cases of the condition fulfillment) for which, say, the ratio of losses to profits equal to 40/60 would convince you that there is a statistical advantage?
Matemat, you're shifting my focus. I've written so many letters, but I guess I didn't get the message across. I'll try again. But first I'll answer your question.
First of all, it must be tested on a logically justified hypothesis of a certain behavior of the market in certain circumstances. I didn't speak about a set of indicators, but only about the "hammer" candle and the hypothesis that the hammer is a reversal candle. It is important for the hypothesis to have a logical basis. For example, the hammer is a reversal candle because the market went far down, turned around and came back up during the formation of the candle. Below I will show why the rationale is so important to me. By the way, I am satisfied with the "hammer" testing methodology in the mentioned article up to the very moment where the author concludes that the hammer does not give a significant statistical advantage. In my opinion, we should have put a full stop here and said that the hypothesis "At the appearance of such a candle we should expect the price increase" was rejected.
By the way, the sample size is fine with me. I'm sure it would be enough to apply a 50/50 ratio of losses to profits to show a statistical advantage. But it's not this that worries me, it's the heresy that you obviously misunderstood to call "an honest attempt to create an indicator evaluation system ".
Now in order. To quote one of the main objectives of the author's work: "1) to find a set of statistically valid values (combinations) of indicators that predict further trend movement with a high probability."
Mathematician, I argue that on any price series other than a constant, it is possible to "find a set of statistically valid values (combinations) of indicators that predict further trend movement with a high probability" exactly in the sense that author does it. I agree that this set will squeeze profits out of history with a mixture of indicators, stops and takeaways. Moreover, it is my deep conviction that this set is uncountable. I.e. there are an infinite number of such "statistically valid" values to be found. And the search among the wipes is a drop in an infinite sea where such values can be found. Therefore, the results obtained by the author are negligible and not applicable in practice. The irony is that the results of the article are not statistically valid. The author has found a single set out of an infinite number of similarly negligible ones, and has not confirmed in any way the opposite - the significance of his result. That is why it would be reckless to use the results of the article in practice.
All the author has done is to find out at which outlets the winning tickets of the previous lottery draw were purchased. He claims: "Look, if you bought up all the lottery tickets in TSUM, some of them lost, but another part won considerably more!" So what? Is the idea of buying up lottery tickets in TSUM statistically valid and giving a higher probability of winning? Only to the unsophisticated.
Alas, Matemat, I do not see in the article any "attempt to create a system for evaluating indicators", but only to create a system of how to fish out a single insignificant combination from an infinite set of insignificant combinations.
Note, Matemat, that even worse, such method does not prove in any way that among the infinite set of combinations there is at least one workable one, which will be usable in the future.
I will not be warmed by the fact that I can spend my whole life combining indicators, takes and stops, and find infinitely many combinations suitable only for the past life. Why do I need mountains of rubbish? And this is exactly what the author suggests - to walk up to a mountain of rubbish and take a piece from this mountain. That is all. This article offers nothing else. How do you determine that the piece found is a valuable grain? Does the author's system allow you to separate the grains from the chaff? Alas, it does not. How could you call it a "scoring system"? And where did you see it as an attempt?
I believe that there is a much better chance of winning the banal lottery than there is of finding a workable combination. Consider the odds for, on the one hand, a clearly existing win against a finite number of options in the lottery, and, on the other hand, a statistical advantage whose existence is in doubt, against an infinite number of "winning" combinations for the past. That's why I take a different approach - to have a logically valid hypothesis, which is what should be tested. Agree, even if I use the method of the author and get one combination, which proves its profitability on a demo account, would I sleep well and put it on a real account until I logically prove, why exactly this combination is so miraculous? And will you, Matemat, bet money on the author's results?
I'll summarise my opinion:
1. 1. The article suggests the method of selecting the only combination from an infinite set of combinations that would show profit on history.
The author doesn't give any proof of the importance of the search method, the set where these combinations are searched for, or the result. And he does not point out the lack of substantiation and the insignificance of the results.
3. I do not find in the article even "an attempt to create a system of evaluation of indicators", but only an apparatus for removing pieces of rubbish from mountains of rubbish.
4. The untrained person may not know that they are being offered a negligible result out of an infinite number of equally negligible results.
Vita, I thought we were past the "hammer" - but thanks anyway for clarifying the position.
And in the last post I was talking about this particular indicator system. In this case the wording of the problem misses the exit condition, because in your general formulation the problem doesn't make sense. Perhaps it is as follows: "Stop - 20, Take - 5"? I'm betting that the m.o. of the trade will not be more than 1 pip, and the frequency of profitable trades - a maximum of 80-85%, not 99. Shall we check it?
Vita, I thought we were past the "hammer" - but thanks anyway for clarifying the position.
And in the last post I was talking about this particular indicator system. In this case the wording of the problem misses the exit condition, because in your general formulation the problem doesn't make sense. Perhaps it is as follows: "Stop - 20, Take - 5"? I'm betting that the m.o. of the trade will not be more than 1 pip, and the frequency of profitable trades - a maximum of 80-85%, not 99. Shall we check it?
Quite right, missing the exit condition. I believe that when you are dreaming about a grail, you should not go down to "unimportant" details, because details immediately indicate the absence of statistical advantages and fall to the ground.
I didn't attribute ma6>ma12 ma12>ma24 ma24>ma48 ma48>ma240 set of indicators to mine, so I didn't understand your reference "the set of indicators you mentioned". Pardon.
Let's not check the set on purpose, for I know there's nothing behind this formula but naive intuition. Neither the first ma6>ma12... nor the second ma6[i]>ma6[i+1]... author's examples are not a clear upward movement. Even without looking at the graph, and having an idea about the nature of moving averages, we can safely say that the formula will catch the rare long rises upwards (not always from the beginning of the rise) and frequent hill tops, including the beginning of descents from the hills, on which the entire advantage of the formula will be compensated to zero. The frequency of profitable trades will not be as optimistic as you suggest. I'll have the nerve to say, that for the pound the frequency of profitable trades will be worse than 20/(20+5+3)=0.71, and for the euro - worse than 20/(20+5+2)=0.74 when applied to ma6>ma12 ..., because there is no stat advantage at all.
Think for a second, why is it necessary that ma48>ma240 is observed? Is there a sensible answer that catching 5 pips should go with ma48>ma240? If it is vice versa - ma48<ma240, 5 pips will be caught worse, because the trend seems to have reversed? Then it turns out that we possess the trend indicator, which is not true, as ma48>ma240 also has no statistical advantage. I doubt that behind this formula lies comprehension of market behaviour or homework with moving averages.
When I asked if anyone had any statistics, I wanted to work out for myself what milestones local people have reached in such a hopeless case. So it turned out that no one who spoke out was interested in statistics. I have been given formulas with the suggestion that I should do the statistics myself. I get the impression that people are in the clouds and are comfortable there with their formulas. Tell me, Matemat, do you have a formula for 5 pips per day with statistical advantage?
Hi all!
First of all I would like to say that discourse is normal, as Elder and Billy said, - "The stock market game only attracts smart enough people and so on..."
You've already done a lot of debating here, it's all very well spelled out.....
I'd like an indicator for pips of 10 pips :))) What for?
---------------------------------------------------------------------
If you take a deposit of 100$.
lot 1.0 (1 tick = $1, 1:100)
-------------------------------
If you trade 4 currencies, doing 10 pips each, you can earn 40$ (40% of deposit) every day, in couple of days you will get your deposit paid back. And all is fine,
you gradually build up your deposit, leverage and so on.
Everything is just "PISCES" - How wonderful!!!!
BUT IN THE FIRST POSTS IT SAID: "IF we take 5 pips and a stop loss of 15, at 50*50, we lose. . . . " - PISES,
What to do????
----------------------------------
// I want to note I'm only playing on DEMO, intraday, I'm a beginner... //
And here's what I was thinking...
Our savior is a trend. i.e. 10 pips+SWOPs+commissions = open only the most powerful trend (which occurs on average 0-2 times within a day (powerful trend + rollback)).
Not in any case, not in corridors, flats, ranges. We cannot afford small corrections!
By the way, what else I wanted to say is why this is real.
On average within a day price fluctuates between 60 and 120 pips approximately, depending on the currency the trend can go 50-60 pips. So if you go into a trend, I think 10 pips can be taken out, that's the most important thing. (the main rule of the trend, like the next bar is bigger than the previous one, so I think the stop loss will not work either)
I have a few rules (some of them are the same as the ones I've already given you, others I've added):
1 you can only play one currency once when there is a strong trend. You may have it once a day or not at all. (It is a mystery how to understand that this is a powerful trend and how to enter it before it has had time to make 10% of moves). This is where forex put me on my haunches. :(
2 If we have a powerful trend, 50 pips,
let's say,
- I keep 10 pips to confirm the move.
- 20 take away
- 20 goes further without me, or for especially greedy ones, we move with stop losses.
again, everything is just fine, but where is the powerful trend ?????