Forex Factory (https://www.forexfactory.com/forum.php)
-   Trading Systems (https://www.forexfactory.com/forumdisplay.php?f=71)
-   -   Simple Mean Reversion (https://www.forexfactory.com/showthread.php?t=743125)

VEEFX Dec 1, 2018 3:01am | Post# 421

{quote} Yes, I'm pretty sure. What you see on the chart is a stress test with maximum leverage 500:1 ! And the system still survives. This means that with lower leverage this system is very safe.
Just sharing my experience... I spent months developing and optimizing a multi-strategy, multi-pair, multi-tf system on 500:1 leverage using offshore broker which was extremely profitable (by design) only during fat tails (diversified divergence from mean instead of reversion to mean). When I moved to 50 leverage, it balked because it limits the scope of diversification and the way I calculate my position sizing based on FreeMargin. What you said is true only if you have a very large capital base to compensate for the leverage and have no impact on diversification across the intended multiple pairs of choosing. Just sharing my own experience which turned out to be quite disappointing in the end. I would also say (not sure if this is applicable in your design).... I would question the merit of long term success of any strategy that is dependent or sensitive to Account Balance or Account Leverage or Position Sizing in general. To me, these are the tell tale signs of a heavily optimized strategy to curve-fit the market movements from past.
I really hope you have found the gold brother. I never doubted your intelligence :-)

Masterrmind Dec 1, 2018 10:09am | Post# 422

{quote} Yes, I'm pretty sure. What you see on the chart is a stress test with maximum leverage 500:1 ! And the system still survives. This means that with lower leverage this system is very safe.
Certainly the rest of the trade analytics would be useful here instead of just showing a single chart.

But its your show.

Masterrmind.............

simjo Dec 1, 2018 10:30am | Post# 423

THINK! PATTERNS! MARKET MODELS! TOTAL NUMBER OF VARIATIONS??? CYCLES! TIME DIVIDERS! RELATIONSHIPS BETWEEN TIME AND PRICE! PRICE OVER TIME! PRICE MOVING OVER TIME! (Not time over price!) TIME IS THE KEY! TIME DETERMINES PRICE! (Not the other way) NUMBER OF MARKET MODELS BASED ON FIXED TIME CYCLES? HOW MANY? 3 ! 1. HORIZONTAL RANGE (Many variations) 2. RANGE WITH DIRECTIONAL DRIFT (Many variations) 3. SMOOTH DIRECTIONAL TREND (Many variations) WHAT ARE THE PERCENTAGES FOR EACH MODEL TYPE? 1 - DOMINATES! 2 - APPEARS FROM TIME TO TIME. 3 - VERY RARE!...
Secrete of No 6 is market conditions?

1 High volatility
2 Low volatility
3 Trending up
4 Trending down
5 Ranging
6 Transition condition

market made up on these conditions.

zigler Dec 1, 2018 12:36pm | Post# 424

"1, 2, 4, 5, 7, 8 1 * 1 = 1 2 * 2 = 4 4 * 4 = 16 1 + 6 = 7 5 * 5 = 25 2 + 5 = 7 7 * 7 = 49 4 + 9 = 13 1 + 3 = 4 8 * 8 = 64 6 + 4 = 10 1 + 0 = 1 1 * 2 * 4 * 5 * 7 * 8 = 2240 2 + 2 + 4 + 0 = 8 NO 3 6 9 ??? THINK OUTSIDE THE BOX! 1 + 4 + 7 + 7 + 4 + 1 = 24 2 + 4 = 6 1 + 4 + 16 + 25 + 49 + 64 = 159 1 + 5 + 9 = 15 1 + 5 = 6 6 + 6 = 12 1 + 2 = 3 6 * 6 = 36 3 + 6 = 9 3 6 9 HMMM! DOWN THE RABBIT HOLE....... NO TURNING BACK!
Yeah, and all you actually need is:

1 - Buy, 0 - Sell (up/down, high/low). That be it...

Everything else is meaningless.. Sell high, buy low, zoom f*g out for better results, cause on 5min you really don't know
where you are... Kinda Photoshop trading, if it's high on screen sell, if its low, buy, do this on 30+ pairs and you got profit...
Simplify as possible, not otherwise...

Jagg Dec 4, 2018 7:01am | Post# 425

3 Attachment(s)
I stopped posting the weekly results a few weeks ago.... For anyone interested here are the monthly results (and all time result from july until friday last week). More or less good profits were generated only for the indices CFD (Dax and Dow) - forex doesn't look that good. 9990 = my variant of SMR with 4H MA filter, BE stop, no fixed timed 25%/50% exits but "candlestick jumping stop loss" - see above for more details 9993 = original method from post 1 with only BE...
Missed the all time results from magic number 9993 - so here is the last peace of my gathered results

(for better comparison again the all time results from magic number 9990)
Click to Enlarge

Name: 2018-11-23 Results ALL TIME 2 - from July 2018 until November 20181123 Simple Mean Reversion EA .png
Size: 14 KB
Click to Enlarge

Name: 2018-11-23 Results ALL TIME - from July 2018 until November 20181123 Simple Mean Reversion EA 99.png
Size: 10 KB

mav3n Dec 4, 2018 9:42pm | Post# 426

...... Here you can test one of my new tools: Drawdown logger. {file} This indicator will log your account drawdown in real time with perfect accuracy! ............
Dear Alpha,
Is it possible to have the drawdown for each pair opened? Thanks in advanced

alphaomega Dec 5, 2018 1:17am | Post# 427

{quote} Dear Alpha, Is it possible to have the drawdown for each pair opened? Thanks in advanced
Yes, it's possible, but not with this indicator.
This indicator records the difference between the balance and equity taking into account all open positions. When you don't have open positions it does not record anything.

mav3n Dec 5, 2018 1:39am | Post# 428

{quote} Yes, it's possible, but not with this indicator. This indicator records the difference between the balance and equity taking into account all open positions. When you don't have open positions it does not record anything.
Dear Alpha, thanks for the quick reply, looking forward for the 'each pair drawdown' Indicator .

Thanks anyway for the info.

Chake Dec 5, 2018 11:51am | Post# 429

AO, with your DD logger i can see one version of this system i made, sometimes reaches 50% DD intraday... But always comes out ahead eventually.
How much intraday DD do you experience in your system if I may ask? And how much would you consider too much?

alphaomega Dec 5, 2018 12:13pm | Post# 430

AO, with your DD logger i can see one version of this system i made sometimes reaches 50% DD intraday... But always comes out ahead eventually. How much intraday DD do you experience in your system if I may ask? And how much would you consider too much?
There is no such thing as too much drawdown on account basis. All depends on how you manage your capital and how much you are willing to lose.
All about risk:reward.

If You don't understand what I mean, you can check one of my old posts on this topic. This is how I manage my capital.
You can apply this system no matter how large or small your capital is.

Chake Dec 5, 2018 12:54pm | Post# 431

{quote} There is no such thing as too much drawdown on account basis. All depends on how you manage your capital and how much you are willing to lose. All about risk:reward. If You don't understand what I mean, you can check one of my old posts on this topic. This is how I manage my capital. You can apply this system no matter how large or small your capital is.
great post, thanks!

alphaomega Dec 6, 2018 8:35pm | Post# 432

1 Attachment(s)
Do you want to know WHY this system works??
Here is the answer in the form of nicely visualized statistical data.
This is the price distribution (close prices on M5) from the last 200 days on EURUSD.
Do you notice the cluster in the first half of the day??
BOOM!

Click to Enlarge

Name: EURUSDM5.png
Size: 65 KB

zerocash Dec 9, 2018 7:15am | Post# 433

Do you do your statistical analysis in MT4?
Have you changed scale in/out or do you still add to losses on this EA?

Nice EA thanks for all your posts, not just this one.

alphaomega Dec 11, 2018 6:33am | Post# 434

1 Attachment(s)
Can you see the pattern?
Click to Enlarge

Name: The Secret Order.png
Size: 464 KB

alphaomega Dec 11, 2018 7:19am | Post# 435

If the market moves from point A to point B in a straight line over time T, then what is the actual length of the movement from A to B ?
If the market moves up from point A to point B and down from B to point C, then what is the actual length of the movement from A to C ?
If the market moves down from point A to point B and up from B to C and down from C to D, then what is the actual length of the movement from A to D?
If the market moves........

Yes, that's right. The more it moves, the longer the length becomes. Or in other words, the frequency of directional change + the amplitude determines the total length.

Now, here comes the real question!
Where the total length of price movement is longer? During range, or during trend? And why this even matters?
What happens when you know precisely the total length of price movement for all time frames and all markets all the time?

DavidRP Dec 11, 2018 7:59am | Post# 436

If the market moves from point A to point B in a straight line over time T, then what is the actual length of the movement from A to B ? If the market moves up from point A to point B and down from B to point C, then what is the actual length of the movement from A to C ? If the market moves down from point A to point B and up from B to C and down from C to D, then what is the actual length of the movement from A to D? If the market moves........ Yes, that's right. The more it moves, the longer the length becomes. Or in other words, the frequency...
well I have more questions. How are A,B,C,D defined? Local max /mins?

alphaomega Dec 11, 2018 8:11am | Post# 437

1 Attachment(s)
{quote} well I have more questions. How are A,B,C,D defined? Local max /mins?
The best way is to use fractals from different time frames! I simply apply the standard formula with 5 bars. You can apply this formula even down to the tick level.

BTW Just look at the EURUSD for example. How much movement there is! The total length is enormous! Thousands and thousands of pips every week!

Click to Enlarge

Name: EURUSDM5 Fractals.png
Size: 132 KB

AntiCre Dec 11, 2018 9:45am | Post# 438

What happens when you know precisely the total length of price movement for all time frames and all markets all the time?
Reminds me somehow on th π and river movement analogy. However, if you use the tick chart to get the best estimate of the market fluctuations, you'll find that π is very different from 3.141 ...

So, if we would precisly know the total length of price movements for a particular time and time frame, we could somehow predict how price has to act. Example: We know that price has to move X ticks within 1 hour. Now we waited 55 minutes but price just went X - 142 ticks. Thus, price has to move more (either in an oscillating = ranging or in a straight line manner) to reach X within the remaining 5 minutes ...?!

DavidRP Dec 11, 2018 11:18am | Post# 439

{quote} Reminds me somehow on th π and river movement analogy. However, if you use the tick chart to get the best estimate of the market fluctuations, you'll find that π is very different from 3.141 ... So, if we would precisly know the total length of price movements for a particular time and time frame, we could somehow predict how price has to act. Example: We know that price has to move X ticks within 1 hour. Now we waited 55 minutes but price just went X - 142 ticks. Thus, price...
Yep that's fine but again, you need direction in order to profit from the move. At the start of the day I know a pair is going to move >60 pips, but I need direction to get the pips..

AntiCre Dec 11, 2018 11:49am | Post# 440

{quote} Yep that's fine but again, you need direction in order to profit from the move. At the start of the day I know a pair is going to move >60 pips, but I need direction to get the pips..
Yes - I was never able to do something useful with it, especially since its variation is too high in short-term applications ...


© Forex Factory