Simple Mean Reversion | Page 7 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

A Simple Mean Reversion Strategy 392 replies

Pairs Trading: Reversion to the Mean 22 replies

Forward test of my new mean reversion strategy 11 replies

Synthetic hedges, cointegration, mean reversion and similar stuff 419 replies

Cointegration, Synthetic hedges, mean reversion in R, Tech Thread 0 replies

Attachments: Simple Mean Reversion
Exit Attachments

Simple Mean Reversion

  • #121
  • Apr 4, 2018 4:49pm Apr 4, 2018 4:49pm
Quoting alphaomega
{quote} You answered your own questions! Of course that we need minimum range. This is part of the strategy but I left the details for next posts. (I can't write long posts now) After all we try to capture only 50%(max) of the range during the first half of the day. So for example if we have range 5 pips it does not make sense to open position because the maximum potential profit is only 2.5 pips. This is why in the second system layer I use sqrt(t) to determine optimal entry points. I have this calculation in graphical/visual form as well. I will...
Can you share this file as well?
  • #122
  • Apr 21, 2018 11:40am Apr 21, 2018 11:40am
been testing this vigorously for a few days and it always results in account blowing up at some point. unless the creator of the system has a way to mitigate a one way market this system will never work. i think you can take some risk off with tailoring around news events, but that alone will not cut it.

the only option is to cap the positions it takes i presume and that could be tested if it was an input in the EA..can you add that to the EA ??

looks like thread is dead so i'll leave it at that.
supply & demand is a farce !
  • #123
  • Apr 21, 2018 1:12pm Apr 21, 2018 1:12pm
Hi Alphaomega, if i'm not wrong you have replaced you profile picture right? last time is Alpha and Omega sign. And i know you are a genius when you have a discussion with CP or maybe when CP refer to your post, i don't remember exactly.

Okay i'm not a genius like you, so i think it will be much more time for me to grasp your lesson, but i'll try. Please be patient..

Thanks for sharing, btw.
  • #124
  • Apr 24, 2018 7:38pm Apr 24, 2018 7:38pm
Quoting bumb
been testing this vigorously for a few days and it always results in account blowing up at some point. unless the creator of the system has a way to mitigate a one way market this system will never work. i think you can take some risk off with tailoring around news events, but that alone will not cut it. the only option is to cap the positions it takes i presume and that could be tested if it was an input in the EA..can you add that to the EA ?? looks like thread is dead so i'll leave it at that.
The system works. (The concept is proven)
Read the basics again.
You have more than you know!

The thread is not dead. Just on pause because I'm too busy with other projects.
Some are related to this system!

Attached Image (click to enlarge)
Click to Enlarge Name: algobot1.png Size: 71 KB

Attached Image (click to enlarge)
Click to Enlarge Name: sdx1.png Size: 154 KB
  • #125
  • Edited 6:04am Apr 25, 2018 4:24am | Edited 6:04am
Quoting Masterrmind
{quote} That is for contrarian trades what about the 'standard' type trades. Quoting directly from Post 1 the Thread Op states: "

3. Position Close.

  1. If we have open position and the current time is before the mid cycle (12:00), then we look for correction to the mid range. (50% retracement from the high)
  2. If the current time is after 12.00, then we look for 25% retracement from the high.


In the second layer we are going to add more precise and continuous calculation using nonlinear formula. For now it's not necessary. Regardless...

I'm still struggling with the second bullet "after 12:00 looking for a 25% retracement from the high" Now, since positions may only be entered before 12:00, this 25% retracement from a post-12:00 high may be well in the negative, right?

EDIT: To illustrate a bit, here's a screenshot from a day gone bad for the system:
Provided there would have been an entry level for the last bar before mid-cycle, there was never a 25% retrace from high in the second half of the cycle (colored in red). Even if there would have been, the exit would have been above the last short entry.
Attached Image (click to enlarge)
Click to Enlarge Name: drs.png Size: 74 KB
  • #126
  • Apr 25, 2018 7:01am Apr 25, 2018 7:01am
Quoting mfeldt
{quote} I'm still struggling with the second bullet "after 12:00 looking for a 25% retracement from the high" Now, since positions may only be entered before 12:00, this 25% retracement from a post-12:00 high may be well in the negative, right? EDIT: To illustrate a bit, here's a screenshot from a day gone bad for the system: Provided there would have been an entry level for the last bar before mid-cycle, there was never a 25% retrace from high in the second half of the cycle (colored in red). Even if there would have been, the exit would have been...

In my opinion, the system works like this :


1) You showed an entry at 1500 hrs. This is forbidden according to the system rules. No entries after 1200 hrs.

2) Now,let us suppose, this entry was before 1200 hrs. This entry is NOT the highest high at that moment,
but must also obey the condition :

ShortEntry >= Level_4, where

Level_4 = Open+(atr/8)*sqrt(hour+1)

Afterwards, the trend was against you. Well, in that case you run into your stoploss. Which is defined as :

Level_6 = open+(atr/3)*sqrt(hour+1);

where atr = daily average range over 20 days. For EURJPY this results in around 100 pips at 1500 hrs.

You calculate it at the moment, you take the trade. From then on, it remains a fixed Stop Loss price.

3) The 25 % retracement level for your Profit Target must also be calculated at the moment, you take the trade.
So, your intended exit is never above the entry price, but always below it. (In case of an initial short trade).

4) if you show a chart, please show the price axis also. That makes it easier to discuss targets and stoplosses.
  • #127
  • Apr 25, 2018 8:58am Apr 25, 2018 8:58am
Quoting Riverstar
{quote} In my opinion, the system works like this : 1) You showed an entry at 1500 hrs. This is forbidden according to the system rules. No entries after 1200 hrs.
Er, no - 15:00 is arbitrary - it's the civil time according to the broker's time zone. The system rules exclude entries after mid-cycle, which the OP calls 12:00.
You will notice that my day starts at 03:00, the forbidden zone at 15:00 accordingly. The supposed entry is in the 14:45 bar.

Quote
2) Now,let us suppose, this entry was before 1200 hrs. This entry is NOT the highest high at that moment, but must also obey the condition : ShortEntry >= Level_4, where Level_4 = Open+(atr/8)*sqrt(hour+1) Afterwards, the trend was against you. Well, in that case you run into your stoploss. Which is defined as : Level_6 = open+(atr/3)*sqrt(hour+1); where atr = daily average range over 20 days. For EURJPY this results in around 100 pips at 1500 hrs. You calculate it at the moment, you take the trade. From then on, it remains a fixed Stop Loss price....

That's why I said *if* there was an appropriate entry level. In the first post there was no "Level 4", and in particular no SL mentioned, these came later with the sqrt(t) rules.
I agree that introducing SL will change the picture completely, in particular whole "look for a 25% retracement" scenario becomes largely redundant.

However, as long as there's no SL but only the 25%r and close-at-end-of-cycle thing, there will be plenty of days as the one shown where the system runs into deep problems.


Quoting Riverstar
You calculate it at the moment, you take the trade. From then on, it remains a fixed Stop Loss price. 3) The 25 % retracement level for your Profit Target must also be calculated at the moment, you take the trade. So, your intended exit is never above the entry price, but always below it. (In case of an initial short trade).
Now that's something interesting where I think the rules are actually not clearly defined by the OP. Is the "take profit at 50% retracement between 00:00 and 12:00" always referring to the current mean level, or to the mean level at the moment the trade was entered?

To me the first seems to be the case, as shown in post #30 . The close for the two sells is indicated on the current yellow line, sell1 is thus closing in loss, only sell2 is in profit.
  • #128
  • Apr 25, 2018 1:13pm Apr 25, 2018 1:13pm
I hate reading tea leaves and that is exactly what we are doing here so far. As the system was explained and which I've read several times this system reminds me of a chase betting system. You can win 50 games in a row (per series 7 game chase) and lose just one chase and most if not all the winning games are gone.

As the way the system is now and the very limited EA inputs the OP has provided I would recommend NOT TO TRADE this system right now. Hopefully the OP will get serious with this thread because I have some ideas and input I would like to provide, but I will wait until system is revealed..

So far everything looks good accept dealing with fat tail days and sometimes weeks... could blow up your account very easily.
supply & demand is a farce !
  • #129
  • Apr 26, 2018 2:36am Apr 26, 2018 2:36am
I copied the indicator for Pineview. For the interested: You can play around with it here:

https://www.tradingview.com/chart/EU...-Range-System/

You'll quickly find that, at least as initially described, it has more days with problems than not.
  • #130
  • Apr 26, 2018 10:40am Apr 26, 2018 10:40am
Quoting mfeldt
{quote} I'm still struggling with the second bullet "after 12:00 looking for a 25% retracement from the high" Now, since positions may only be entered before 12:00, this 25% retracement from a post-12:00 high may be well in the negative, right? EDIT: To illustrate a bit, here's a screenshot from a day gone bad for the system: Provided there would have been an entry level for the last bar before mid-cycle, there was never a 25% retrace from high in the second half of the cycle (colored in red). Even if there would have been, the exit would have been...
The screenshot says alot.

This is the typical behavior that is something mean reverting traders attempt to avoid at all costs (DD escalation !).

As I've mentioned many times, there appears to be no hard and fast strategy to overcome this evil DD except a hard SL or close at Time_X.

I was hoping on a more productive and faster discussion with the Thread Op however, it appears at this stage he is far too busy to get involved.

I'll move on for the time being and continue with my own projects but I'll drop by now and again.

Good trading,

Masterrmind............
Master your Mind then Master your Trades
  • #131
  • Apr 26, 2018 1:56pm Apr 26, 2018 1:56pm
Quoting Masterrmind
{quote} The screenshot says alot. This is the typical behavior that is something mean reverting traders attempt to avoid at all costs (DD escalation !). As I've mentioned many times, there appears to be no hard and fast strategy to overcome this evil DD except a hard SL or close at Time_X. I was hoping on a more productive and faster discussion with the Thread Op however, it appears at this stage he is far too busy to get involved. I'll move on for the time being and continue with my own projects but I'll drop by now and again. Good trading, Masterrmind...............
Exactly ! I have always been fascinated by the concept, and have tried several things to try and mediate the huge DD issue it always faces. I trade it but use a very tight SL on trades that go do not retrace much. So far i have not found a real good solid edge with the system. Interesting to see the second installment of the system because i have struggled with DD on this approach... green pips !
supply & demand is a farce !
  • #132
  • Apr 27, 2018 1:38pm Apr 27, 2018 1:38pm
Quoting mfeldt
{quote} I'm still struggling with the second bullet "after 12:00 looking for a 25% retracement from the high" Now, since positions may only be entered before 12:00, this 25% retracement from a post-12:00 high may be well in the negative, right? EDIT: To illustrate a bit, here's a screenshot from a day gone bad for the system: Provided there would have been an entry level for the last bar before mid-cycle, there was never a 25% retrace from high in the second half of the cycle (colored in red). Even if there would have been, the exit would have been...
Would you mind sharing this indicator?

Thank you
  • #133
  • Apr 28, 2018 2:26pm Apr 28, 2018 2:26pm
Quoting Mini7260
{quote} Would you mind sharing this indicator? Thank you
I thought I already did that, but now I notice I shared only the "idea", not the indicator script.
it's hopefully here: https://www.tradingview.com/script/V...-Range-System/
  • #134
  • Edited Apr 29, 2018 8:41am Apr 28, 2018 3:05pm | Edited Apr 29, 2018 8:41am
Quoting alphaomega
{quote} Yep. This is pretty much the only potential flaw of the system. The week spot. The Achilles' heel Every system has a weak spot. That's why the whole development of this system is concentrated on different methods to reduce the impact of these undesirable fat tails. The good news is that almost half of the fat tails are caused by economic data(news). And most of these news are known in advance. They are scheduled...
Deleted to avoid clutter...
Staying in my lane...
  • #135
  • Edited 4:10pm Apr 28, 2018 3:51pm | Edited 4:10pm
Quoting VEEFX
Would you agree that trades taken on divergence alone will yield close to 50:50 win/loss? it's more in the range of 46:54. Would you agree that trades taken on reversion alone will yield close to 50:50?
Just to add ... The global ratio similar to those you mentioned. From my own research I've seen one number occuring over and over again ...

  1. What is the probability that a bar has the same color as its preceding bar?
  2. What is the probability that a ZigZag repaints?
  3. What is the probability that a 5-chunk of candles fulfill the 5-linear subsequence theorem if the previous 5-chunk doesn't fulfill it?

The answer is always close to 52.5 %!

PS: Write a free pdf and offer some courses teaching the "Time-Action Analysis". Live from the fees

  • #136
  • Apr 28, 2018 4:18pm Apr 28, 2018 4:18pm
Quoting AntiCre
{quote} Just to add ... The global ratio similar to those you mentioned. From my own research I've seen one number occuring over and over again ... What is the probability that a bar has the same color as its preceding bar? What is the probability that a ZigZag repaints? What is the probability that a 5-chunk of candles fulfill the 5-linear subsequence theorem if the previous 5-chunk doesn't fulfill it? The answer is always close to 52.5 %!
Thanks AntiCre. Hope you are going well. I know you have a immense aptitude of analytical research mindset in you. I have to agree on your first point. I was being liberal with the 46:54 split giving some room for margin of error. Mine also tells me the real truth in trading that no one wants to accept. There are millions of binary conditions that could occur and no matter what strategy we deploy on which market condition, the probability will almost always fall within the 48%-52% win versus loss. That is why my Challenge to the trading community is still not broken. "A trading strategy that purely relies on price:time series on just one chart (one pair and one timeframe) will not survive more than 3 years". 3 Years is somewhat arbitrary to eliminate the human luck factor and to incorporate discipline in execution with discretionary style. There is not a single automated strategy that can last for 3 yrs in retail forex that is based on one chart one pair one timeframe. It cannot exists because if it were to exists, market will seize to exits and that will be the demise of the entire trading industry (which has started on a false premise and will continue the 'system' of brainwashing retail trading young and old to fall prey on their marketing promise.

Feel free to prove me wrong. Use any tools, objects, indicators, candle formation, price action on a two dimension chart with price over time. There has to be atleast ONE external factor introduced in the strategy for it to succeed. Depending on a price or time derivative is just not enough.

On your second point, I don't even understand the meaning of "repainting" of indicators as I am against using them... even for ATR. They are just illusion causing roadblocks or obstacles.
I will leave the 3rd point for others to comment :-
Staying in my lane...
  • #137
  • Apr 28, 2018 4:27pm Apr 28, 2018 4:27pm
Quoting VEEFX
They are just illusion causing roadblocks or obstacles.
Well, they are measuring tools. A ZigZag is basically designed to identify breakouts. You see how people fool other traders. Moreover, there are only a couple of real indicators. For instance, ZigZags are the same as Stochastics, the same as Semaphors, Cycle Identifier, Williams % R, Donchian Channels, etc ... And the most crazy thing is that they have different standard settings. Who wonders that they give different signals?!
  • #138
  • Apr 28, 2018 5:43pm Apr 28, 2018 5:43pm
Quoting VEEFX
{quote} Hi Alpha, I know you are a smart dude. Based on your research.... Would you agree that any system strategy is doomed for failure in the long run if it is designed to handle only certain market conditions? Example: if a fat tail in your example is a dooms day scenario, no amount of filtered can be applied on price or time or any combination there of to avoid the dooms day. Would you agree that any "optimization" based only on Price is a flawed strategy in of itself? Would you agree that price will always diverge or revert from mean or whatever...

Hallo veefx,
Your comment has been awarded the title of "complete bullshit" and has nothing to do with the original intent of this thread.
Please, do not disturb an ongoing discussion with items that are irrelevant to it.
  • #139
  • Edited 8:07pm Apr 28, 2018 7:46pm | Edited 8:07pm
Quoting VEEFX
{quote}
.......

  1. Would you agree that any system strategy is doomed for failure in the long run if it is designed to handle only certain market conditions? Example: if a fat tail in your example is a dooms day scenario, no amount of filtered can be applied on price or time or any combination there of to avoid the dooms day. Probably Yes. "On a long enough timeline the survival rate for everyone drops to zero"
  2. Would you agree that any "optimization" based only on Price is a flawed strategy in of itself? I don't know.
  3. Would you agree that price will always diverge or revert from mean or whatever level 100% of the time? No. The price does not move most of the time.
  4. Would you agree that trades taken on divergence alone will yield close to 50:50 win/loss? it's more in the range of 46:54. No. (Other factors in play)
  5. Would you agree that trades taken on reversion alone will yield close to 50:50? No. (Other factors in play)
  6. Would you agree that market does not have a contract with Price to stay within a range or to conform with ATR/ADR? Yes and No.
  7. Would you agree that bi-directional trades taken on both diversion and reversion will increase your win/loss ratio? No.
  8. Would you agree that uni-directional trades taken on both diversion and reversion will increase your win/loss ratio? No.
  9. Would you agree that uni-directional trades taken on divergence and countertrend trades taken on reversion will increase your win/loss ratio? No.
  10. Would you agree that unidirectional trades taken on reversion and countertrend trades taken on divergence will increase your win/loss ratio? Maybe.
  11. Would you agree that bidirectional trades taken on both divergence and reversion would increase your win/loss ratio? No.
  12. Would you agree that bidirectional countertrend trades taken on both divergence and reversion would increase your win/loss ratio? No.

  • #140
  • Apr 30, 2018 4:22am Apr 30, 2018 4:22am
Quoting leppozdrav
{quote} You are right. DD almost equals the final return. Increasing the account size is ok and would lower the DD, but it's just a bandage for the strategy or "escaping the problem" in this case. I would focus more on filtering those bad spikes and close them sooner - limit the opening of new orders. So, Alphaomega suggested to play out this strategy on uncorrelated Fx pairs. I don't say, we should just put it on a net of baskets and let it run. Choosing the right Fx pairs is one of the keys here for this to go better. So I did that. Below is a...

How is it possible that your average profitable trade, with a lotsize of 0.01, is $ 11.64 and your losing trade - $15.48
In general, the width of a trade is 25% to 50% of the daily range. That means, with your results, you have a daily range of 232 or something like that.
But in reality the daily range is much smaller. For eurusd the ADR = 80 pips. So, profitable trades should be between $2 and $4 on average.
0 traders viewing now