Similarity: A Practical Approach

The backtest for the "Daily Projection Method" doesn't even need to be done. You can visually see it for the last 2 months using a 1H and Daily separation bars.
I mentioned this in the post about the method:

It was a bullish trend since Bernake's Bullish Bar (200+ pips I think??), therefore:

1. Price above 300EMA, default projection should always be UP DAY.
2. Confidence signs:
   1. Price in the upper channel of Bollinger
   2. Stoch at upper channel of Bollinger
   3. Price tends to be above or in the same angle as ZigZag lines.

3. Retracements
   1. Counter-trend trades (ie shorts) with short-term targets.
   2. Price makes ZZTop and crosses the Bollinger Mid towards the other side of BB (this can take 1-3 days to happen)
   3. Price continues once ZZBot has been confirmed
   4. Alot more volatile if price overall is approaching or nearing the 300EMA which can act as support but that doesn't mean price won't go below the 300EMA to test bearish strength.

If you look at the original thread, I did one week's testing of live calls using a more stripped down eyeball method of looking at the chart.
I think I got 2 days wrong out of 3. One of those days was a draw, ie. Doji with price diff of 0.0004

Recently, I've not had the need - but I always begin the day with a projection.... a mini bias. Yesterday I was UP but I was wrong. That doesn't matter though because you still go through the StochBB method using the 1M/5M/15M resolution and then check in on the 1Hour maybe 3 times a day? So even if you're "wrong" on the daily, you can still be right on the trade.

Hope that helps

Here's the Demo Account, warts and all including the bad trades when I still wasn't sure of the signals and also some experimental trades where I hedge.
You will note that no stop loss is used. Various reasons I won't go into now. Enjoy

I have a pending sell at 1.33365, top of the DS Box on the 1H. This could go higher than that..... hmmm +30?

Took another 44 pips from longs around 1.32920

Interesting market movements are translating into our indicators.

1.33014 is now the top support level of a new dissimilarity zone on the 1H.
Price is above it, so it can only be cleared from above unless a major whipsaw operation occurs.

However, in reaching this price, it resolved a black hole in the previous swing hi/lo that I mentioned earlier today.
So I am wondering whether that "energy" will uncoil itself into higher levels before a massive short. Just thinking out loud.
The 4H chart is screaming for a new ZZTOP that will break 1.3345

I'll continue to post this for those following the BBSTOCH method.

Before I project "Daily Direction" for tomorrow, I noticed this on the Daily TF.
The last ZZTop 1.34149was in the BB!!= Black Hole!

The resulting price action moved away from it, then DS Started and it moved towards it again to resolve it and create a new ZZTop. I'm not sure whether that will be this week or the next, but 1.34149 needs to be resolved.

The 1H chart makes it difficult to project tomorrow's direction.
I want to say it's an UP Day.

However, the 4H shows that a ZZTop is due.
The 1H tf also concurs that a ZZTop is due.

If this eventuates early in the day, then the next likely direction is down. This is what makes the Daily projection difficult.

What I do know is that the Daily tf is projecting a target of 1.34149. A Bottom is needed before this price is reached.
Therefore, a move south is required and could be synch'd to my feeling that a Down Day may eventuate.

The 1H shows that a DS Zone can be cleared at 1.3337. If this is reached during Asian session then a southern move back to 3300 would make a great retrace for the higher target the following day. I'm going to say DOWN DAY.

Thanks Ralome, This reason sounds clear. Is there another way to finish dissimilarity before 24 candles? Any kind of agreement between stoch and price, both in relation to BB? Or dissimilarity candle forces this 24 canldes dissimilarity period?

Another thing I think I have adapted latest indicator version so it can replace rv-2. I hope I did not missed anything. It was rather quick check.

Just a thought......

Can someone code the indicator to actually highlight the dis-sim candle? Might help you guys with visuals, especially when it comes to calculating where price will be one the zone has been painted.

Very Loose Backtest of "Daily Forecast" Method posted a while back:

1. Data taken from MT4 and compiled onto spreadsheet
2. "Up" or "Down" day based on Open being Higher or Lower than Close
3. Sample size taken from 1st Jan 2013 to 1st August 2013
4. Manual entry based on 1H and method/rules from earlier post (sorry too tired to look it up here)
   1. UP day scored as 1
   2. DOWN day scored as 0
   3. Total 130 Up Days against 182 Rows

5. Result: 71.43% overall win/loss ratio
   1. Jan 69.23%
   2. Feb 70.83%
   3. Mar 80.77%
   4. Apr 57.69%
   5. May 77.78%
   6. Jun 64%
   7. Jul 77.78%

Hi, nice effort! just to confirm it is your post 474: http://www.forexfactory.com/showthre...64#post6860764

1st question: different broker starts a new day at different hours. do you think that in fact the beginning hours of Aus/Japan are significant instead of a certain broker's 1st hour?

2nd question: Are below the only 5 rules you used to predict next day's up/down, or you incorporate any additional new considerations?

If PRICE is above 300EMA, the default daily direction will always be UP. Other things to look for to support an UP projection

1. PRICE is also traveling the upper half of the bollinger band
2. PRICE closes the day near a ZZBottom and the next major swing into the new day has to be a TOP.
3. STOCHBB is above the 50 Line
4. Stoch is also travelling the upper half of the bollinger band​

3rd question: for the PRICE closes the day near a ZZBottom and the next major swing into the new day has to be a TOP; both ZZbottom and TOP should happen before tommorrow (within today)? or only the TOP should happen in the beginning of tomorrow?

your test results are simply astounding and incredible. it is an "EDGE" that every trader dreams for. i'll attach an EA that can amplify your deadly accurate daily prediction into ridiculous profit.

The EA's this week performance is here: http://www.mt4i.com/users/charvoke/stats, check the "charts" tab for the equity.
Since you know tommorrow's direction, simply input 1 for buy or -1 for sell in 'SpecPairsBias', and keep 'UserBias' true. The EA will amplify your correct prediction's profit by crazy stacking positions. Please try it for 1 or 2 days(or multiple pairs if you like). I look forward to combining your method reading tomorrow with this EA if interested.

Hi, I have been silently following the all the "similarity" threads, been using Ralome's "StochDifZones" indicator, and desperately trying to get some value out of all of it .

Here are few observations I made so far:

1. Most people seem to forget about one significant information that EURUSDD shared in the original thread (post #3476).
   
   

   Quote

   Disliked

   The orIginal principle states that price MUST return to the level where the dis-similarity was FIRST DETECTED. So, even if price returns there and the zone has not yet appeared (according to Ralome's indicator), the CONDITION IS DONE!!!!!

   
   If you actually apply this rule, you may observe many long-hanging zones to be in fact invalid according to the above principle. Surely, some of the good-looking ones are cleared early too, but still I believe this to be very useful. My hope is that once we account for that, it can significantly improve the ratio of only ~60% good zones rendered by the indi.
2. I tried to analyze the differences between r-v1 and r-v2 indis prepared by EURUSDD, the only one I spotted (besides different default parameters) was use of SMA in r-v1 and EMA in r-v2. I am not sure how important might be.
3. Ralome's indi identify the sim/dissim zone by checking the difference between MA(1) within bb and stoch(300,1,1) with its own bb. The difference must be greater than some value. That value in Ralome's calculations is: 0.0001.
   If you modify this value to let's say(0.001, 0.01, 0.1), then it, as expected, will generate shorter but more frequent dissim periods.
4. Isn't the main principle of the similarity theory the fact that the data must be >~90% similar? If this is true, then what we see on our charts is much less than that. Now playing with what I described in (3) and also not counting dissimilarity periods as dissimilar after they were "cleared" (1), must get us closer to the "90+% must be similar" principle.

Having all of the above in mind and being a programmer,but totally new to MQL4 I took the liberty and attempted to apply some of the above ideas to Ralome's indicator :

1. Close unborn zone and make abortion after price returns to the level where dissimilarity was first detected.
2. Allow to define MA type (parameter: "ma_type").
3. Allow to define the value that differentiates between dissimilarity and similarity.
4. Display the diff in a positive/negative way, to understand "which way" the difference goes.

Please be advised that attached indi is highly experimental, and may be unstable. It contains a lot of comments and commented out code blocks. Any feedback (especially from Ralome) here or priv will be greatly appreciated.

Regards!

Ok 1st off sorry I am asking this...this has been answered a few times already but I can't find it. I did have the formula written down some where but I can't find it either. Help me with this and I'll be mindful to keep it in my dropbox so I won't loose it again...of course unless I forgot the password...

Ok I know how to convert stochs..post 240 anwsered it

Can't remember how to convert ZZ to higher and lower TFs.......using 24,5,3 for 15m.

BBs should be similar also?

Sorry doing a chart "reset".Trying to make things more user friendly and stuff.

Edit thanks V! Just trying to overlay the ZZ on top of each other figure I just have to multiply upwards

Edit lol there is actually a MTF ZZ......

Morning All,

15M has just posted a ZZBot inside BB = Black Hole @ 1.32916
Price may need to return there before making higher.

@cfdtaleong
Use 4500 min as the base. So for the STOCH settings
15M = 4500 / 15 = 300
05M = 4500 / 5 = 900
01H = 4500 / 60 = 75

24 was for the BB. But I don't know if you convert this.
For the ZZ, Ive combined the use of ZZ and Semafor. If they are both similar, the turn is stronger.
The setting for ZZ is always 24. The Semafor, I use default and I find they are in sync in any timeframe!
Dont know why!

Hope this helps!

In case of r-v1 it's forced, yes. And your update is correct, if you changed nothing else besides the MODE_EMA's

You upped the wrong file, this is identical to my v2.30 (in which I accidentally wrote v2.10).

Yet added the code I found in rvs. Before that there was a difference in what rv-2 and rv-2_2_60 show.

Added code:

marked it with comment here.
And associated parameters... not listed there, taken directly from rv-2

i can not see any ma_type defination...

This bit is a feature I've removed. It was still there when Eurusdd made the EMA-mod and posted rv-2.

Right, sorry about that! Rookie mistake . Here it is.

I'd vouch you for the efforts if I could. You've also shown that you've been really reading all this stuff

The idea to abort the zones (yeah this is a sicko thread lol) is great. This will cut a lot of clutter. However observations have shown that price does display certain behaviors at these abortion zones which could be potentially profitable also.

Our biggest grapple with the zones thus far has been when and which behaviors price will exhibit at such zones. I am currently for the idea that they at least have some kind of memory effect. Without some kind of conclusive result, attempts to trade them has been risky at best in my opinion. Vlady's and others work on the ZZ methology (ZZOLOGY!)has however been able to mitigate that into something very workable.

I will when I return to my platform run your indicator in tandem with ZZology and see how the game plays out if we look only at the "virgin" zones.

Thanks! Going through all posts in all threads was a pain. The amount of information is overwhelming and it is hard to dig out the useful stuff from ~300 pages. I started collecting the most useful information starting from post #1472 in the original thread into a text file, but never had time to go back to it, plus I have to go over the first 1471 posts again to search for some more gold .

Anyway, I was hoping we could summarize the Similarity principles here, so we are using as much bits and pieces of information, EURUSDD gave us, as we can. Of course, this is just my opinion others may have done that already. Once and IF I am done with the summary, I will share it here, but it may take some time .

Regards.