The similarity system | Page 598 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Similarity: A Practical Approach 859 replies

Similarity Trading - The ZZBB Method 215 replies

Rainbow similarity 256 replies

Similarity Stuff 17 replies

Directional Similarity 0 replies

Attachments: The similarity system
Exit Attachments

The similarity system

  • #11,941
  • Nov 9, 2014 7:33pm Nov 9, 2014 7:33pm
Quoting kprsa
Hi all, here is the version of the Recurrence Statistic that is lighter on the processor (the big calculation is done once every bar now, even this is an overkill). I have left the labels the same for continuity, even though the label of k is imprecise. Several minor bugs have been fixed. I guess it is best if somebody else gives a version with alerts if required, best not to complicate the basic version too much. I add a pdf that covers the basic usage of the indicator. Happy trading, k {file} {file} {image} Edit: error in text corrected, thanks...
Thank you kprsa
What you do is a reflection of you....what I do is a reflection of me.
  • #11,942
  • Edited 9:44pm Nov 9, 2014 7:55pm | Edited 9:44pm
This is just the FORMAT!!! Not a signal!!


Attached Image

Attached Image
  • #11,943
  • Nov 9, 2014 7:58pm Nov 9, 2014 7:58pm
Quoting kprsa
Hi all, here is the version of the Recurrence Statistic that is lighter on the processor (the big calculation is done once every bar now, even this is an overkill). I have left the labels the same for continuity, even though the label of k is imprecise. Several minor bugs have been fixed. I guess it is best if somebody else gives a version with alerts if required, best not to complicate the basic version too much. I add a pdf that covers the basic usage of the indicator. Happy trading, k {file} {file} {image} Edit: error in text corrected, thanks...
Hi k,

Thank you very much for making your PDF. I suppose I should make one at some point for my NT7 indicators too!

I have some questions about the section "The k business". Maybe it's just my fuzzy Monday morning head so apologies for my misunderstanding:

1) The numbers in this paragraph...

"For example, in Fig 1 we see that the <k> for the end-of-the bar with specified input parameters is 0.0002 = 2 pips. While the probability that the zone will be cleared within specified h_right is 85%, the probability that the price will reach within 6 pips if it is not cleared (3*<k>=3*2 pips) to clearing it is 99.3197% !"

... don't seem to match the numbers in Figure 1 on page 3. Is this a mistake?

2) "Additionally, it gives the probability that the price will retrace within 1*<k>, 2*<k>, 3*<k>."

Presumably 1*<k> means it reaches the edge of the average k zone? And 2*<k> means it reaches the edge of double the average k zone?

3) "Note that this statistics happens in addition to the one of fully cleared zones, so you can have a, say, probability of 80% to clear the zone and, say, 90% probability to come very close to clearing the zone from the cases that did not clear the zone, so the probability of this trade is close to 98% (80%+90%*20%=98%)."

Sorry, I don't understand this calculation. Where does the 20% come from? And what is that 98% the probability of?

Thanks,
NT
You reap what you sow.
  • #11,944
  • Nov 9, 2014 8:00pm Nov 9, 2014 8:00pm
Quoting kprsa
{quote} the logic is that the two numbers can be different. you can and should play with the default values. k
And dont be misled by a post that you may see that said "if markets are fractal then the H should remain the same across time frames".

Fractals are exponential/logarithmic and constrained by the resources in the environment needed to create them. You may expect a storm once a year to come along and double the size of a 6 foot deep 10 foot wide ravine in a few days rain. But you cant multiply by 1000 and expect that every 1000 years a storm will come along and double the size of the Grand canyon (6000 ft deep and 10000 ft wide) by raining a few thousand days. Not enough energy or water on the planet to support that storm. So the H value of grand canyon doubling doesnt scale with size.
In the same manner, General Electric may bring a few billion home to the US and cause a market move that creates TZ. But on a monthly chart who is going to do something that causes 45 bar H transient zones?
  • #11,945
  • Nov 9, 2014 8:05pm Nov 9, 2014 8:05pm
Quoting NorthTrader
{quote} Hi k, Thank you very much for making your PDF. I suppose I should make one at some point for my NT7 indicators too! I have some questions about the section "The k business". Maybe it's just my fuzzy Monday morning head so apologies for my misunderstanding: 1) The numbers in this paragraph... "For example, in Fig 1 we see that the <k> for the end-of-the bar with specified input parameters is 0.0002 = 2 pips. While the probability that the zone will be cleared within specified h_right is 85%, the probability that the price will reach within...
i think i'm the only stupid around here who can't put all toghether to trade)) but no worrys keep drayning the banks boys
  • #11,946
  • Edited 9:50pm Nov 9, 2014 8:59pm | Edited 9:50pm
went short aud/usd
using Kiads BB TZ system @ 8668 tp lower bb 8636 sl 8701 RR 1:2

the big white line is 360 SMA so trend is down for me
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 66 KB
There has never been a statue erected in honor of a critic
  • #11,947
  • Nov 9, 2014 9:17pm Nov 9, 2014 9:17pm
Quoting pip.monkey
went short aud/usd using Kiads BB TZ system @ 8668 tp lower bb 8636 {image}
Careful the mid bar ptz
I want more brain power!
  • #11,948
  • Edited 10:31pm Nov 9, 2014 9:51pm | Edited 10:31pm
hi guys,

i finished mod xixi's tz indicator last weekend.
i added k-filter value and stats for it. it should look like this.

http://i.imgur.com/pGl9f1O.png

if you want solid rectangle, set back = true, fill = true and line style = solid.

http://i.imgur.com/6UVpvEx.png

i dont know what to explain
you just play around with it to find your h and k to increase your probability.

if it has bug, let me know.

hope you guy find it useful.

have a green pips week

zelo

added: % recurrent is when price return and touch k-zone > 1 time.

edit: fix bugs
Attached File(s)
File Type: mq4 xixi_tz_1.8.mq4   19 KB | 994 downloads
When you lose, I'm the loudest who laugh at your face )
  • #11,949
  • Nov 9, 2014 9:51pm Nov 9, 2014 9:51pm
Quoting kprsa
Hi all, here is the version of the Recurrence Statistic that is lighter on the processor (the big calculation is done once every bar now, even this is an overkill). I have left the labels the same for continuity, even though the label of k is imprecise. Several minor bugs have been fixed. I guess it is best if somebody else gives a version with alerts if required, best not to complicate the basic version too much. I add a pdf that covers the basic usage of the indicator. Happy trading, k {file} {file} {image} Edit: error in text corrected, thanks...
To Kprsa many thanks K great work ,really appreciated .
Regards orid peter
  • #11,950
  • Nov 9, 2014 9:52pm Nov 9, 2014 9:52pm
USD is rising above Zero Level. Gaining strength.

Time to sell Gold?

Lets see.
Attached Image (click to enlarge)
Click to Enlarge Name: gold sell..PNG Size: 44 KB
Traders gain some knowledge, win, their emotions kick in, and they die.
  • #11,951
  • Nov 9, 2014 9:57pm Nov 9, 2014 9:57pm
Quoting Zelo
hi guys, i finished mod xixi's tz indicator last weekend. i added k-filter value and stats for it. it should look like this. {image} if you want solid rectangle, set back = true, fill = true and line style = solid. {image} i dont know what to explain you just play around with it to find your h and k to increase your probability. if it has bug, let me know. hope you guy find it useful. have a green pips week zelo {file}

Here comes another Gem. Luv u Bro
Traders gain some knowledge, win, their emotions kick in, and they die.
  • #11,952
  • Nov 9, 2014 10:02pm Nov 9, 2014 10:02pm
congratulations thanks for your hard work
  • #11,953
  • Nov 9, 2014 10:09pm Nov 9, 2014 10:09pm
Quoting NorthTrader
{quote} Hi k, Thank you very much for making your PDF. I suppose I should make one at some point for my NT7 indicators too! I have some questions about the section "The k business". Maybe it's just my fuzzy Monday morning head so apologies for my misunderstanding: 1) The numbers in this paragraph... "For example, in Fig 1 we see that the <k> for the end-of-the bar with specified input parameters is 0.0002 = 2 pips. While the probability that the zone will be cleared within specified h_right is 85%, the probability that the price will reach within...
1) you are correct, it should be 99.4197. it is an (obvious) typo.
2) within <k> means that the price has retraced to less than <k> from clearing the zone.
3) 20=100-80. 98% is the probability of the trade expecting a very narrow transient zone. there are two sets of stats shown by the indicator, one is for the cleared problem bar, another for the uncleared. if one wants to apply this to a trade, one should take into acount both possibilities.

k
  • #11,954
  • Nov 9, 2014 10:13pm Nov 9, 2014 10:13pm
me again

when Eursdd's Sub Sequence, TZ and Similarity come into play on EURUSD recently.

http://i.imgur.com/tgsi67v.png

http://i.imgur.com/pvhjEVp.png

When you lose, I'm the loudest who laugh at your face )
  • #11,955
  • Nov 9, 2014 10:23pm Nov 9, 2014 10:23pm
Quoting Zelo
if you want solid rectangle, set back = true, fill = true and line style = solid.
dude, this setting is not working, lol, build mt4 745, thanks....
  • #11,956
  • Nov 9, 2014 10:34pm Nov 9, 2014 10:34pm
Quoting paladim
{quote} dude, this setting is not working, lol, build mt4 745, thanks....
i have edited the code.
it should work now

thank you for your report.

you may want to combine it with kprsa's indicator to find your h and k.
When you lose, I'm the loudest who laugh at your face )
  • #11,957
  • Nov 9, 2014 10:39pm Nov 9, 2014 10:39pm
Quoting Zelo
{quote} i have edited the code. it should work now thank you for your report. you may want to combine it with kprsa's indicator to find your h and k.
Awesome! It's working now...

and 10s for the combination tips! hehe.
  • #11,958
  • Nov 9, 2014 10:47pm Nov 9, 2014 10:47pm
Quoting Barbaresco
{quote} Normally I focus on only the numbers and stats - tonight I've been looking at charts with the data in mind... 1. TZ's generally oscillate Top/Bottom 2. Mid TZ's should not be traded against 3. The first PTZ of a direction is very unlikely to become a TZ So - if we know (or at least suspect) the likelihood of a PTZ becoming a TZ increases exponentially with time, then, once 50% of the time has elapsed, it might be reasonable to expect a PTZ to be a TZ. If it is a TZ, given observations 1-3 above, I think we can develop some rules for high...
Some more M1 data:

1. Assuming any PTZ that is older than 50% of your h (chosen for 3%) will become a TZ, and trading away from this "TZ", with a TP of a PTZ in the opposite direction, and an SL at the assumed TZ, yielded approx. 61% W:L, with less than favorable R:R. Net loser, but getting closer.

2. Trading away from a confirmed TZ ( put SL here) if and only if a PTZ has not yet been made in the opposite direction (TP when it makes the PTZ) yielded approx. 72% W:L, also with less than favorable R:R. Net it comes out a little better than breakeven. I have not yet tested how forcing a better R:R changes the W:L.

I was hoping to not have to filter TZ's to make them work, but it seems to be heading that way.
To test next:
2a) 2 from above but with better R:R - can filter the entry based on computed R:R at entry (decreases number of trades), or can simply increase the TP and/or decrease the SL to fit desired R:R at entry
3. Confirmed Top or Bottom TZ followed by a PTZ in the same direction - trade against this - thinking of scaling in and TP'ing all orders when an opposite PTZ is made - have yet to code, but EU looked like only 14 trades last week - once coded will look back to find worst case scenarios to figure risk and SL

**note - some tests have shown profitable in higher timeframes, but I am focusing only on M1 for now
  • #11,959
  • Nov 9, 2014 10:54pm Nov 9, 2014 10:54pm
Quoting Zelo
hi guys, i finished mod xixi's tz indicator last weekend. i added k-filter value and stats for it. it should look like this. {image} if you want solid rectangle, set back = true, fill = true and line style = solid. {image} i dont know what to explain you just play around with it to find your h and k to increase your probability. if it has bug, let me know. hope you guy find it useful. have a green pips week zelo added: % recurrent is when price return and touch k-zone > 1 time. edit: fix bugs {file}
good work pro
  • #11,960
  • Nov 9, 2014 11:40pm Nov 9, 2014 11:40pm
Quoting kprsa
1) you are correct, it should be 99.4197. it is an (obvious) typo.
Actually I meant all the numbers in that paragraph (0.0002, 2, 85%, 6, 99.3197%). None of them match the fractal stats in Figure 1. All typos?

Quoting kprsa
2) within <k> means that the price has retraced to less than <k> from clearing the zone.
Got it. Thanks.

Quoting kprsa
3) 20=100-80. 98% is the probability of the trade expecting a very narrow transient zone. there are two sets of stats shown by the indicator, one is for the cleared problem bar, another for the uncleared. if one wants to apply this to a trade, one should take into acount both possibilities.
I suppose it depends on what you mean by "very narrow". I can't see how this probability can be higher than the 90% one for "very close to clearing the zone". I wonder if this is how some traders are quoting such high trade probabilities when in reality the frequency calculations yield lower ones?
You reap what you sow.
16 traders viewing now