Explore Real (Statistical) Market Edges | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

TOP 100 best edges on forex charts: post good edges here 116 replies

Statistical mean of the market [quant corner] 191 replies

Statistical Validation on trading 24 replies

Attachments: Explore Real (Statistical) Market Edges
Exit Attachments

Explore Real (Statistical) Market Edges

  • #1
  • First Post: Edited Sep 13, 2021 1:38pm Jun 11, 2020 2:39am | Edited Sep 13, 2021 1:38pm
Hello Friends,

This threat is dead... No need to read
  • #2
  • Jun 11, 2020 5:14am Jun 11, 2020 5:14am
Before to add any indicator I'd suggest you to apply a time filter to find best hours to trade. I suspect London opening hours will give the best performance.
  • #3
  • Jun 11, 2020 6:17am Jun 11, 2020 6:17am
Naked Explorer Test Curve

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_raw_1.PNG Size: 18 KB


Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_raw.PNG Size: 69 KB


Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_raw_2.PNG Size: 24 KB


Result: EXPLORER shows nice down curve. Losses are caused of Spread. The Balance Curve is not straigth, because I used not 99% or 90% Backtest but only with Controllpoints. Thats enough for a first test.
  • #4
  • Edited 2:24pm Jun 11, 2020 6:22am | Edited 2:24pm
EURUSD, H1, EMA 200
01.01.2019 - 11.06.2020
Rules: Original EXPLORER plus BUY only IF ASK is higher than EMA200 and SELL only IF BID is lower than EMA200

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_MA_200_1.PNG Size: 26 KB


Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_MA_200_3.PNG Size: 26 KB


Result: 50% less trades, but same losses. Drawdown is higher with EMA200 rule than without.

Conclusion: EMA200 is not respected by price. If it would respect EMA200, in the long run (over 1000 Trades a small edge would be visible).
  • #5
  • Jun 11, 2020 6:26am Jun 11, 2020 6:26am
Quoting BermeoFX
Before to add any indicator I'd suggest you to apply a time filter to find best hours to trade. I suspect London opening hours will give the best performance.
Thank you BermeoFX, good point. I will include time factor in next test for the EMA200 test.
  • #6
  • Jun 11, 2020 10:48am Jun 11, 2020 10:48am
There is no repeatable edge in the markets. The most important thing is to make your losses smaller than your gains. Some situation if you do better research than the market you can have an edge over that situation like finding that the euro is 15% overvalued, etc. But they don't repeat, you need to do another research tomorrow. You can mine the data all you like, you won't find anything. I once found a pattern that predicted what the market will do (60% of the time). But with time , it disappeared. As the market changes and morphs , old patterns dissappear and new ones emerge. Some people are wired for that, i wasn't. So l took an approach where I don't have to do constant research, all I need is strong discipline. Hope it helps.
When you see it, BET big.
  • #7
  • Jun 11, 2020 2:35pm Jun 11, 2020 2:35pm
Quoting Rennaissance
There is no repeatable edge in the markets. The most important thing is to make your losses smaller than your gains. Some situation if you do better research than the market you can have an edge over that situation like finding that the euro is 15% overvalued, etc. But they don't repeat, you need to do another research tomorrow. You can mine the data all you like, you won't find anything. I once found a pattern that predicted what the market will do (60% of the time). But with time , it disappeared. As the market changes and morphs , old patterns...
Thank you for reply. Its really appreciated.

If there would be no edge in the market, than I would stop trading. Because than trading would be only gambling. Gambler always lose money in long term. In my eyes there is no professional and consistent profitable gambler. If a gambler is profitable on long run, than he has an edge. But than its not gambling. Its trading. For example good poker player or Casino players very often now about their statistical edge. If you say that the most important is to make more money than to lose, than also this we could test. I will do this in next test. I will play around with different SL and TP. How does an bigger TP improve the outcome? It should all be visible and possible to test.
  • #8
  • Edited 2:50pm Jun 11, 2020 2:40pm | Edited 2:50pm
Quoting kleybenny
EURUSD, H1, EMA 200 01.01.2019 - 11.06.2020 Rules: Original EXPLORER plus BUY only IF ASK is higher than EMA200 and SELL only IF BID is lower than EMA200 {image} {image} Result: 50% less trades, but same losses. Drawdown is higher with EMA200 rule than without. Conclusion: EMA200 is not respected by price. If it would respect EMA200, in the long run (over 1000 Trades a small edge would be visible).
Interesting fact: I just run a optimization for the EMA200 test. Same rules but only with optimization for the EMA length. See pic:

Attached Image


The outcome was:

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_EMA_optimization_2.PNG Size: 7 KB


Only one EMA value seems to be profitable enough to end in tiny profit after 1700 trades: EMA 130.

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_EMA_optimization.PNG Size: 67 KB


After this I testes AUDUSD h1 for the same time span and with the same rules. Once without EMA and once with EMA 130. The outcome was very poor. I run an optimization and found that for AUDUSD h1 the "magic" EMA value was 600. I changed the time span to a longer time span of 4 Years backtest and: again the result was very poor.

This brings me to the point I mentioned in my first post: NO CURVE FITTING!

The slightly better result for EMA 130 with EURUSD h1 and 600 with AUDUSD h1 was only because we choose the EMA which fitted perfectly to the history. That was why the results were better.

My conclusion for Moving Averages: Moving Averages are lagging indicators. They are only a picture of the past. A picture of history. It has no impact to the future. Wheter a 200EMA, 100EMA, 50EMA, 75EMA, 21EM, what so ever has anything to do with predicting the future price movement. At least a single moving average on a single time frame. Later on I will test the behavior of more than one indicator and in more than only one time frame. But for now my conclusion is clear: EMAs dont work when they are used as a stand alone indicator on a chart.

PLEASE PROVE ME WRONG. OR TELL ME WHERE MY ERROR IN REASONING IS. I WOULD APPRECIATE IT!
  • #9
  • Jun 11, 2020 3:02pm Jun 11, 2020 3:02pm
Quoting BermeoFX
Before to add any indicator I'd suggest you to apply a time filter to find best hours to trade. I suspect London opening hours will give the best performance.
I applied the time filter. I add a filter which allows EA to only open orders in London market hours. The outcome is interesting. First picture shows EMA200 in London Market hours and second picture shows EMA130 in London Market hours.

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_EMA_200_London_Market.PNG Size: 63 KB


Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_EMA_130_London_Market.PNG Size: 72 KB


Both test showed: The results are slightly better in London Market hours for EURUSD. It seems that trading while LONDON MARKET HOURS PROVIDE US WITH A REALY EDGE WHEN USING EMAs. IT WILL BE INTERESTING HOW IT IS WITH OTHER INDICATORS OR OTHER PATTERNS.

At least for now I will let the Market Hours Filter in.
  • #10
  • Jun 11, 2020 3:11pm Jun 11, 2020 3:11pm
The next test I will do is testing the edge of bigger SL in comparison to TP. First I will double my TP to 40 and let SL at 20. Than I will run optimization to see the potential of RR ratio. But again: Optimizations are only curve fitting! After optimization results, I always need to check if the outcome was a real optimization which works always or if it was only true for the given moment. My first feeling says that the RR will have a big impact to the trades also. But lets prove.
  • #11
  • Edited 3:36pm Jun 11, 2020 3:18pm | Edited 3:36pm
RR TEST: For this Test, I changed the backtest to every tick model (using m1 candle data, it is still not 99.99% but good enough for the moment). Thats important, since TP and SL is not the same anymore and fast movements could distort the outcome...

EURUSD, H1, EMA 130, LONDON MARKET, RR: 20 pip SL102 pips TP (1:5)
01.01.2019 - 11.06.2020

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_EMA_130_London_Market_RR_1.PNG Size: 20 KB


EURUSD, H1, EMA 200, LONDON MARKET, RR: 20 pip SL102 pips TP (1:5)
01.01.2019 - 11.06.2020

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSD_h1_EMA_200_London_Market_RR_1.PNG Size: 26 KB


Conclusion? What is your point of view?
  • #12
  • Jun 11, 2020 4:28pm Jun 11, 2020 4:28pm
Hello kleybenny,
maybe our minds work in a similar way (tending to use only half of our brain) ...
Some time ago, reflecting on what you say in the introduction, doubts have struck me ...
Perhaps in the "field" of "logic" the battle is lost for the human being, because there are machines, that have "more advantage", and if have all the money in the world would buy the best machine and the best programmer (but not is the case of the retail trader) ...
Also, there are tons of people with way more intellect than me ...
However, our nature drives us again and again hoping to get different results ...
The "logic" gives the human being a false sense of security ...
why false? because security ends when something escapes logic, and that happens every day many times ...
Maybe using our "full" brain, logic and feeling, help a little ...
A few candles and a few TFs can offer much, much more information than any indicator I have tried ...

regards,
wish you a good way
Newbie trader
  • #13
  • Jun 12, 2020 2:39am Jun 12, 2020 2:39am
Quoting fex
Hello kleybenny, maybe our minds work in a similar way (tending to use only half of our brain) ... Some time ago, reflecting on what you say in the introduction, doubts have struck me ... Perhaps in the "field" of "logic" the battle is lost for the human being, because there are machines, that have "more advantage", and if have all the money in the world would buy the best machine and the best programmer (but not is the case of the retail trader) ... Also, there are tons of people with way more intellect than me ... However, our nature drives us...
Hi Fex, thank you very much for your contribution. I understand what you mean. That could be also an outcome of this test. I like to see what is statistical provable and what not. If indicators, price action behaviors strategies and so on would not have a statistical edge, than I would need to go into another direction. More into the "feeling" or "interpretation" way. Thinks which you could not describe, but what you can learn only with experience. Same as driving a car.... But than I would try to skip all the misleading indicators and strategies which fools me on this way of trading with false hope. Lets see. I´m still quite confident that we can define things in market which provide us a real edge....
  • #14
  • Jun 12, 2020 2:49am Jun 12, 2020 2:49am
Things I consider to test next:

  1. Candle Stick Patterns (Engulfing Candle, V formations, Pinbars...)
  2. Indicators (RSI, Stochastic, Volume?)
  3. ?

I like to hear your opinion.

  • #15
  • Jun 12, 2020 2:55am Jun 12, 2020 2:55am
One thing came into my mind when thinking about Moving Averages. Maybe its not the position of the moving average in relation to the price. Maybe its the steepness or angle of Moving Averages which gives us an edge. I will test this later on....
  • #16
  • Jun 12, 2020 3:33am Jun 12, 2020 3:33am
Hi, may be this treath could help: https://www.forexfactory.com/showthread.php?t=907889

regards
Newbie trader
  • #17
  • Jun 12, 2020 4:29am Jun 12, 2020 4:29am
Good. Just ignore comments like "markets are unpredictable, this works, this doesn't"... without proof.

From personal experience, my first and last "half serious" attempt at using statistics to forecast the future didn't work that well: after almost 2 years of live forwardtesting I ended up with a +10%, which is too low for my taste.

Now, is it because that system just crashed ? is it because I was using too tight stops and slippage killed potential profits ? or I just failed to optimize the parameters ?


I don't know... I took a break from this stuff but it's always good to read about other people experiences.
  • #18
  • Sep 6, 2020 2:15pm Sep 6, 2020 2:15pm
Quoting kleybenny
One thing came into my mind when thinking about Moving Averages. Maybe its not the position of the moving average in relation to the price. Maybe its the steepness or angle of Moving Averages which gives us an edge. I will test this later on....
May I suggest you to read this, this and this posts.
Nice thread by the way. Subscribed.
There is no try
  • #19
  • Sep 7, 2020 10:31am Sep 7, 2020 10:31am
Kleybenny

The thread thus far is limited to trying to identify an edge from the use of technical indicators. However, there is of course always fundamental analysis at your disposal, albeit much tougher to model.

Regarding the use of MAs, one option would be to consider the reversion to the mean - the price move before price reverts back to the mean.

Nice thread - will keep an eye on what you are doing.
  • #20
  • Sep 7, 2020 10:49am Sep 7, 2020 10:49am
If you try a TMA - TRAINGLE MOVING AVERAGE SET AT 60 BY 120 - MULTIPLIER 5.8 TO 8 YOU SHOULD FIND ON A 15 MINUTE CHART IT WILL BE IN THE 80% RANGE ON GBPUSD AND OTHER CURRENCIES WITH A 150 POINT PLUS RANGE. ITS PROBABLY THE MOST ACCURATE MOVING AVAERAGES. ATTACHED A CHART BELOW
0 traders viewing now