MMLCv16 | Page 4 | Forex Factory

Options

Go to
Permalinks
Print Thread
Attachments: MMLCv16
Exit Attachments

MMLCv16

  • #61
  • Jul 18, 2016 7:05am Jul 18, 2016 7:05am
Quoting trader2925
Discovered some errors in excel.Posting Corrected stats. Edit: I had a lingering doubt that something is wrong about L10 stats but since rel verified ok i ignored the instinct.lol. Before it was 300 something for OBS and around 340 for L1O.Now it is around 300 something for both ,L1O and OBS.( for Up and down that is.) {image}
Yup these look much nicer.
  • #62
  • Jul 18, 2016 11:15pm Jul 18, 2016 11:15pm
Quoting Relativity
{quote} Yup these look much nicer.
yea,but still DJ->? has UR-UT numbers pretty wide.Is that okey?what numbers you are getting for DJ->? for period jan2016 to jun2016?
  • #63
  • Jul 19, 2016 1:40am Jul 19, 2016 1:40am
Quoting trader2925
{quote} yea,but still DJ->? has UR-UT numbers pretty wide.Is that okey?what numbers you are getting for DJ->? for period jan2016 to jun2016?
The rest are really ok so I have no issue with this. Nothing can be perfect but we do our best to find the norm based off the law of large numbers.
  • #64
  • Edited 4:47pm Jul 19, 2016 2:05pm | Edited 4:47pm
Quoting Relativity
{quote} The rest are really ok so I have no issue with this. Nothing can be perfect but we do our best to find the norm based off the law of large numbers.
Thanks to LG,i discovered an error in my code.If trigger happen on an outside bar and there is no further close is same direction beyond ma then swing turn is not registered.There are 5 instances of this error which mean 10 swings were missed.But even after correction L1O stays same and there is no change in DJ->?.t:

edit:Thankfully there was no code error.It was just a settings error.Min bars for swings was set to 1 so setting it to zero made the missing (trigger bar only) swings appear:G
  • #65
  • Jul 26, 2016 8:31am Jul 26, 2016 8:31am
The Concept of Balance and Purpose of Optimization (part 2 of 2)

Fortunately, this is a short section, but it would only be successful if part 1 is correctly done. I think it should be ok for the rest who needed more time, before I move onto B2i.

Get the key zone levels based off x(OBC) to x(L1O) data. To get the basic idea of what these zones are and how to get them, read these posts from MMLCv12 :
http://www.forexfactory.com/showthre...19#post8583019
http://www.forexfactory.com/showthre...64#post8585364

... and attempt to get the following results :

a : R(OBC) to R/J/T(L1O)
Attached File(s)
File Type: xls R(OBC) to x(L1O).xls   46 KB | 434 downloads

a1-Minimal Retracement Zone : 0.0 to 34.99
a2-Retracement Success Zone : 35.0 to 94.99
a3-Retracement Failure Zone : 95.0 to 109.99
a4-Breakout Zone : 109.99 to 129.99

d : J(OBC) to R/J/T(L1O)
Attached File(s)
File Type: xls J(OBC) to x(L1O).xls   33 KB | 438 downloads

b1-Minimal Retracement Zone : 0.0 to 34.99
b2-Retracement Success Zone : 35.0 to 94.99
b3-Retracement Failure Zone : 95.0 to 109.99
b4-Breakout Zone : 109.99 to 129.99

c : T(OBC) to R/J/T(L1O)
Attached File(s)
File Type: xls T(OBC) to x(L1O).xls   73 KB | 388 downloads

c1-Minimal Retracement Zone : 0.0 to 19.99
c2-Retracement Success Zone : 20.0 to 59.99
c3-Retracement Failure Zone : 60.0 to 69.99
c4-Breakout Zone : 70.0 to 84.99

These zones are more 'tighter' compared to MMLCv12, especially for Retracement Success Zone. This is because of the Basic Optimization and Level 1 stacking up. The frequency distribution of the levels here compared to MMLCv12, at least looks to me much more improved.

Then plot these levels onto the chart and you should get something like this :
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 55 KB
  • #66
  • Edited 3:29pm Jul 28, 2016 1:31pm | Edited 3:29pm
I've done the retrace type (a) for the extreme's, see the picture below

Attached Image (click to enlarge)
Click to Enlarge Name: Image1.png Size: 131 KB


Before I'll proceed with the other types, can you confirm that the results are OK?

Can there be found an edge with only the above results?

I suppose with this data that we gonna use a scalping strategy?


Indicator with levels as suggested
Attached Image (click to enlarge)
Click to Enlarge Name: Image2.png Size: 21 KB


Thx in advance,
Tom
  • #67
  • Jul 29, 2016 1:10am Jul 29, 2016 1:10am
Quoting WannaLearn
I've done the retrace type (a) for the extreme's, see the picture below {image} Before I'll proceed with the other types, can you confirm that the results are OK?
Results look great.

Quoting WannaLearn
Can there be found an edge with only the above results?
Yes, there are already edges found with the above results. I won't explain this here but its about the same idea compared to MMLCv12.

Quoting WannaLearn
I suppose with this data that we gonna use a scalping strategy? Indicator with levels as suggested {image} Thx in advance, Tom
I wont make the decision how on what kind of strategy one would be using. The levels do indicate statistically the direction one should be trading at any given time. The actual strategy, it is up to you. For me, I do both scalping and swing trading.
  • #68
  • Jul 29, 2016 5:03am Jul 29, 2016 5:03am
Quoting Relativity
{quote} Results look great. {quote} Yes, there are already edges found with the above results. I won't explain this here but its about the same idea compared to MMLCv12. {quote} I wont make the decision how on what kind of strategy one would be using. The levels do indicate statistically the direction one should be trading at any given time. The actual strategy, it is up to you. For me, I do both scalping and swing trading.

I had already red the v12 thread, but for me the edge was never be clear for the moment.

Do you advice that I try to search for an "tradable" edge first?
Or wait for the B2i box...I'm also very curious to that part.

T.
  • #69
  • Jul 29, 2016 8:09am Jul 29, 2016 8:09am
Quoting WannaLearn
{quote} I had already red the v12 thread, but for me the edge was never be clear for the moment. Do you advice that I try to search for an "tradable" edge first? Or wait for the B2i box...I'm also very curious to that part. T.
I am posting something up soon to help you out, which should answer most of your questions.
  • #70
  • Edited 8:48am Jul 29, 2016 8:12am | Edited 8:48am
How to Trade Using Swings

Always consider the line direction of the current swing as our primary interest; will the swing stop soon OR continue to run?

Again, reading my friend LG's thoughts would be helpful : https://lgtrades.wordpress.com/2015/...-summary-2015/. Here I would explain in more detail from my perspective.

#1
Lets start with the fundamental of all swings; a UR swing. The bell curve results of what happens when swing is in R, particularly Retracement Success Zone, tells us two things;
a- that it is some kind of support where price has the potential to stop here like a cushion because of the previous swing
b- or at least struggle to fight and break through resistance that is against its direction.
Do note when we talk about support and resistance here, its all relative to line direction of swing.

#2
This means one can attempt to go long for this case, positive development would mean price 'bounces' and goes up. We can generally expect that price will travel the length of Minimal Retracement Zone for longing in this case.

#3
Here, of course it is natural to question, hey we can try to also go short here and hope to see positive development for shorting. This means we can attempt to do some range trading here; exit the long to take profit and start to short. Attempting to see length of Minimal Retracement Zone, but for shorting this time around.

#4
However price cannot stay like this forever in range. This is where one needs to understand that further positive development for longs here, implies negative development for shorts. If any previous short trades from #3 doesn't work out and price reaches to this stage, its time to take the loss and look for long trades. This is where knowing where Retracement Failure Zone is would be helpful, which also implies a 2nd opportunity for entering long.

#5
One can say the Breakout Zone in this case has 4 purposes;
a- profit target if one has a long entry in #4
b- if one has any previous short trades from #3, this is your last chance to take the loss before price continues to run up. Generally I don't recommend doing this and take the loss earlier at #4 instead.
c- a possible opportunity to short again, hoping price gets pushed down back and the up swing ends up being a UJ swing, then price travels Minimal Retracement Zone for shorting. J swings still have a bell curve, but not a very strong one compared to R swings. With a weaker bell curve, I generally avoid taking shorts and wait out to look for longs, which leads to the next subpoint
d- the last chance to enter long, before price takes a run as a UT swing. T swings have a flat frequency distribution, which means in some sense, we will never know where it will end, but isn't that the point of 'letting your winning trades run'? This is where the strongest and also most profitable edge exists.

#6
We also know T swings will eventually stop as it cannot last forever, and we can try to trade Minimal Retracement Zone for shorts. The larger the length of T swing, the wider Minimal Retracement Zone for shorts is for taking. B2i will attempt to optimize this. In fact, there are a lot of things one can do to optimize each of these steps from #1 to #6. I would not post further 'what to do and how to do it' (other than just one example; the B2i), but once one gets the general gist of 'what to do', the 'how to do it' portion should be more obvious to you. And you would become more purposeful in seeking new edges and improving existing edges. This itself IMO is the most important edge for a sustainable trading career.

Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 34 KB
  • #71
  • Jul 29, 2016 9:05am Jul 29, 2016 9:05am
Done coding for L10.Many bugs corrected, Only 1 issue remain(levels plotting till trigger bar only).How do you guys code levels for all swings.i have used loop for this and it slows the system down considerably.Dont know if this is software specific issue or coding specific issue ?Will post L1O FD stats tomorrow.

http://i66.tinypic.com/xgelhg.jpg
  • #72
  • Jul 29, 2016 9:10am Jul 29, 2016 9:10am
Quoting trader2925
Done coding for L10.Many bugs corrected, Only 1 issue remain(levels plotting till trigger bar only).How do you guys code levels for all swings.i have used loop for this and it slows the system down considerably.Dont know if this is software specific issue or coding specific issue ?Will post L1O levels tomorrow. {image}
A simple idea if you are up to it...
Use some kind of a on/off switch. Boolean.
For practical purposes / when trading on weekdays, just show the levels for current and previous swings.
For historical purposes / when studying on weekends, show for all.
  • #73
  • Jul 29, 2016 1:45pm Jul 29, 2016 1:45pm
Quoting Relativity
{quote} A simple idea if you are up to it... Use some kind of a on/off switch. Boolean. For practical purposes / when trading on weekdays, just show the levels for current and previous swings. For historical purposes / when studying on weekends, show for all.
Yeah plotting for specific bars only ..... i have already implemented that code.This problem seems to be specific to coding with loops specially when dealing with large data set.Anyways thanks for suggestion.
  • #74
  • Jul 30, 2016 6:01am Jul 30, 2016 6:01am
Hi rel., I will try to give my thoughts about using swings (for the first points)

Quoting Relativity
#1 Lets start with the fundamental of all swings; a UR swing. The bell curve results of what happens when swing is in R, particularly Retracement Success Zone, tells us two things; a- that it is some kind of support where price has the potential to stop here like a cushion because of the previous swing b- or at least struggle to fight and break through resistance that is against its direction. Do note when we talk about support and resistance here, its all relative to line direction of swing.
So in this case the previous swing is a UR swing and the current swing (that isn't finished) is a DR swing @ this point of time.

I also can see the bell curve for the current DR swings, see picture below:
Attached Image (click to enlarge)
Click to Enlarge Name: Image1.png Size: 11 KB

So I agree with the next quote: If the current swing will end as retrace swing there is a high probability that the low/high will be in the "succes retracement zone"
(The catch here is the first part of the quote)

BUT @ that moment we don't know yet that current swing will end as retrace?
Attached Image (click to enlarge)
Click to Enlarge Name: Image2.png Size: 82 KB


In fact from the stats there is more than 50% chance that an extremum will reach 115%.

Quoting Relativity
#2 This means one can attempt to go long for this case, positive development would mean price 'bounces' and goes up. We can generally expect that price will travel the length of Minimal Retracement Zone for longing in this case.
1) I suppose we need to find the 'bounce' at some lower timeframe (M1/M5) because otherwise the min. retr. zone is already reached?
2) At the point we think to see a 'bounce', the next swing isn't yet confirmed (M30), otherwise there is no space left towards min. retr. zone?
3) Most of the swings will have several 'bounces' before the real 'bounce', but even the fake 'bounces' will reach min. retr. zone very often?
4) Space towards min. retr. zone is most of the times very small (3-6 pips)

I hope my comments are a little bit clear.
  • #75
  • Jul 30, 2016 7:27am Jul 30, 2016 7:27am
[quote=WannaLearn;9058295] 3) Most of the swings will have several 'bounces' before the real 'bounce', but even the fake 'bounces' will reach min. retr. zone very often?

There are many fake rejection also with reject swing later turning into trend swing.This is perhaps due to using close of the extreme bar (and not that of next trigger bar) for determining swing types.
  • #76
  • Edited 10:11am Jul 30, 2016 9:56am | Edited 10:11am
Quoting WannaLearn
BUT @ that moment we don't know yet that current swing will end as retrace?
You are right. We don't know. Nobody does. But we are taking advantage of the fact that the chance of it ending there is reasonable, or at least struggling to push through, that itself is an edge already. We are talking about probabilities here, not certainties. but at least we know where and when, rather than none at all. Its very good start already.

I've always used the basic MMLC for 'defensive responsive' style of trading. To move it into a type of 'offensive anticipatory' type of tool, one has to work on more things (I've posted in the other thread that I will not reveal them, but see below).

Quoting WannaLearn
{image} In fact from the stats there is more than 50% chance that an extremum will reach 115%.
Yes and no. Thats not how I would read this. If you noticed, 115% is near to 110%, which is the L1O for R. 110% serves as an important balance point in this situation. Hence why we perform basic optimization and level 1 optimization. This means you will know the following ;
1-The odds of where price ends before it reaches 110% and after it reaches 110%, are now technically about 50/50.
2-Of how price behaves before it reaches 110% and after it reaches 110%, are very different.
3-The balance point for R, T and J are different.
This is very important information, but IMO #2 is the most important and I will not explain this further because it is actually related to multiple timeframes..

Quoting WannaLearn
1) I suppose we need to find the 'bounce' at some lower timeframe (M1/M5) because otherwise the min. retr. zone is already reached?
Yes, looking at lower TF is one of the answers. But I would not discuss about this (about multiple timeframes) on this thread. Its another huge topic together which I had marked it as 'Do not disclose'.

Quoting WannaLearn
2) At the point we think to see a 'bounce', the next swing isn't yet confirmed (M30), otherwise there is no space left towards min. retr. zone?
3) Most of the swings will have several 'bounces' before the real 'bounce', but even the fake 'bounces' will reach min. retr. zone very often?
You can refer to LG's blog post regarding this : "This may be have something to do with minimum profit taking at certain levels."

And this is where I would ask you to look at the swing constructions instructions again ;
Quoting Relativity
a-Downswing run = candle closes below LWMA High line
Ensure swing end point is always the lowest extreme before the next new upswing turn happens.
Mark and keep these LWMA High line data points, hide it later.
......
c-Upswing run = candle closes above LWMA Low line
Ensure swing end point is always the highest extreme before the next new downswing turn happens.
Mark and keep these LWMA Low line data points.
.....
c-Hide the LWMA High line data points and LWMA Low line data points.
Why did I ask to keep the LWMA High and Low line data points, even though we would hide them (visually) later? If you still kept the data points, you can turn them on and off... and start finding some answers to these 2 questions. This is beyond what I wish to discuss on this thread (again, about multiple timeframes).

Quoting WannaLearn
4) Space towards min. retr. zone is most of the times very small (3-6 pips)
It will be especially when the previous swing itself has to build up further, or price has low volatility during the period, or it would also be instrument dependent (I trade EURJPY). Thats all fine, really. Not much we can do, but at least we know when not to trade.


Quoting WannaLearn
I hope my comments are a little bit clear.
I understand some things are going to be frustrating, but look, you are asking all the right questions, similar ones I had asked myself when I developed MMLC over time. Do note that I will reach a point where I will refuse to answer, not because I am irritated, but because you might later reach a point where I have to say 'I will not disclose'. This is where all the real hard work begins; to solidify the basic edges found here further to strengthen your own version of MMLC and for your own trading... on your own terms, not mine. All I can do is to give you some basics, but well done to you for the breakthrough in thought process. It should lead you to somewhere good I want others to make up their own conclusions from here, assuming they had done most of the example I've laid out here.
  • #77
  • Jul 30, 2016 10:16am Jul 30, 2016 10:16am
Quoting trader2925
Quoting WannaLearn
3) Most of the swings will have several 'bounces' before the real 'bounce', but even the fake 'bounces' will reach min. retr. zone very often?
There are many fake rejection also with reject swing later turning into trend swing.This is perhaps due to using close of the extreme bar (and not that of next trigger bar) for determining swing types.
Yes, good observation. This is where you are talking about signal testing. This is important but would only make sense after we are fully clear about the basics, of what Level 1 Optimization can offer us. Hence I've decided not to discuss about signal testing in this thread, but to expound as much as I can on the basics only. Its too much for one person (me) to cover. However, you can discuss this on your own with others.
  • #78
  • Jul 30, 2016 2:40pm Jul 30, 2016 2:40pm
Quoting Relativity
Yes, good observation. This is where you are talking about signal testing.
Thanks rel.This is a weakness of current formula according to me.I will definately explore this after B2i.

Quoting Relativity
Do note that I will reach a point where I will refuse to answer, not because I am irritated, but because you might later reach a point where I have to say 'I will not disclose'.
But you have said you won't respond to stupid questions.How one can be sure why you are not answering.....like you havn't answered my last few pm
and i am still wondering what is the 'stupid' in/about that idea/stats.lol.Now i am hesitating to show more 'ideas' to you through PM.
On MTF analysis I have to say I have been using mmlc12 for mtf ( went crazy following 1m to 1M lol ) and i have noticed 'something' about all swings stats.But had no idea what to do further with the observation.Through this thread i think i may have found a key..Lots of things to do after this thread is complete.Like you said " This is where all the real hard work begins; get your own version of MMLC".
  • #79
  • Jul 30, 2016 3:06pm Jul 30, 2016 3:06pm
Quoting trader2925
{quote} Thanks rel.This is a weakness of current formula according to me.I will definately explore this after B2i. {quote} But you have said you won't respond to stupid questions.How one can be sure why you are not answering.....like you havn't answered my last few pm and i am still wondering what is the 'stupid' in/about that idea/stats.lol.Now i am hesitating to show more 'ideas' to you through PM. On MTF analysis I have to say I have been using mmlc12 for mtf ( went crazy following 1m to 1M lol ) and i have noticed 'something' about all swings...
I think we better program our own MTF_MMLC_v16 indicator....
(Mine was limited for M30, so that's on my priority list now)

Maybe we can better wait for the additional components until we have all the basic information.
But both of us, we have difficulties with waiting...not a good trading skill
  • #80
  • Edited 7:44pm Jul 30, 2016 7:13pm | Edited 7:44pm
Quoting trader2925
{quote} Thanks rel.This is a weakness of current formula according to me.I will definately explore this after B2i. {quote}
Not quite a weakness actually. I placed this 'issue' under signal testing for a few reasons.
1-I had done some comparative stats regarding the standard wick price VS changed wick price if there are inside candles after the extreme is made. The difference isn't much to be excited about, so I decided to stick with standard wick price.
2-What we had always measured as you can see are 'point to point' statistics. As for how does one analyse what happens in the middle, its really another thread itself altogether should I want to post it. But at least I hope you can see that with all this, one can objectively start breaking down how price behaves and analyse each part with ease.
So no, not a weakness, but just more to explore at least to me. I didn't talk about this in MMLCv12 and v16, so it looks like a weakness. I did explore this on my own end and found a few things worth noting, but its not a high priority on my list to continue checking.

Quoting trader2925
But you have said you won't respond to stupid questions.How one can be sure why you are not answering.....like you havn't answered my last few pm and i am still wondering what is the 'stupid' in/about that idea/stats.lol.
I understand english is not your first language based from our PMs. No, I think you don't ask 'stupid questions', more of just 'curious questions'. I decided not to answer them because I decided one thing in advanced; I am no one's guru.

Its not practical for me to answer every single question that comes up to me; I have a bad habit of answering too much in depth (you can see that in the threads lol). And then, parts of 'my secret' can leak out. Alas, then there's no point anymore, because if one can look at the most successful traders, they had earned their keep the hard way. They have all the right to decided what to do with. Most decide not to reveal it. It took me time to earn mine, but I decided to share just a small part of it here. I had to be very selective with what I share. Small and selective enough to help another person see something and I am glad I did.

So to maintain some sanity for myself, some distance is necessary and also helpful for all. I hope you understand my intentions.

Quoting trader2925
Now i am hesitating to show more 'ideas' to you through PM.
Chill lol, there's no need to keep showing almost all the ideas you have in mind to me through PM. Again, as a means to keep things neat and again maintain some sanity for myself, I would only exchange lots of ideas like this back and fro only if there is an agreement between two parties. At this stage, no, I am not open for that kind of exchange.

What I do suggest is this; keep a notebook of all the ideas you have, or perhaps even start your own thread. Take your time to refine them, weed out the weak ones, continue to work on the strong ones, then 'show to the world' just one idea that really works very well after all the testing you had done. Then thats when actual exchange and discussion becomes much more worthy and fruitful.

This is why all my previous threads exist (e.g. Rel's lab and Rel's forge)


Quoting trader2925
On MTF analysis I have to say I have been using mmlc12 for mtf ( went crazy following 1m to 1M lol ) and i have noticed 'something' about all swings stats.But had no idea what to do further with the observation.Through this thread i think i may have found a key..Lots of things to do after this thread is complete.Like you said " This is where all the real hard work begins; get your own version of MMLC".
Going crazy with 1m to 1M? Yes I did that too, and yes I did notice something as well. Well done. And it took me time to properly implement the observation. A long long time. I had that observation during the B2i thread by iDoubleStoch; thats about 1-2 years ago? It was he who hinted to me about it, after I posted a particular conclusion (you can check the B2i thread to find it). Its only about recently I completed the implementation of that one idea. But it is very well worth it.

Yup, go find your own answers to the things you want to look for. On your own time, just pop by this thread for the final section. The thread is more or less about to be completed, after I post about B2i next week.
1 trader viewing now