New historical Forex models | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Markov Models and Forex 33 replies

Anyone used ARMA + GARCH models in forex trading? 14 replies

Historical chart to track historical news events. 0 replies

Risk Models 0 replies

MT4 VWB Models 1-4 8 replies

Attachments: New historical Forex models
Exit Attachments

New historical Forex models

  • #1
  • First Post: Edited Jun 22, 2016 4:21am Feb 27, 2016 10:28am | Edited Jun 22, 2016 4:21am
Hi traders and analysts.

I've been investing and trading for a long time. But I always worry about when I cannot take all the movement entirety. I think you know how it happens. You've opened a sell position, closed your 100 pips, but the market went on. Greed says: "Come on, let's make more money". Many traders have blown up accounts because of it. Long time i was thinking how to take away all market movement and not just part of it. My colleagues, the same traders as I, argue that market is repeated. Namely, it's composed of a unlimited iterations number. The main thing to find this way.

That was my obsession more than six months ago. After that, together with partner hired a staff of 12 analysts. Their main task is in analyzing of current market and predicting future movement based on the past. Simple examples are double top/double bottom or model head and shoulders, or the idea of Elliott Wave theory.

Below I'll give examples of the models and show how they were fulfilled in history.
But there is one problem I think you can help me with. Probability obtaining of positive result 50/50. Other words, we can get the same result, or a totally different. Maybe you'll have any ideas in course of communication that will help me and you to build up our strategies on new models.
  • #2
  • Feb 27, 2016 1:47pm Feb 27, 2016 1:47pm
Its just ridiculous. You hired 12 analysts and you analyzed the double tops and head and shoulders. The problem is you havent analyzed the 20/50 sma crossover. And also the rsi, you know - above 50 buy, below 50 sell. That should get you on the right path.
  • #3
  • Feb 28, 2016 1:14am Feb 28, 2016 1:14am
Matt, I think the right username you should use is Math, instead!

No irony here. One way of making more pips is to start placing limit orders. The rate always bounce back enough. Even in a trend. These limit orders lock down your profit anyway and now you can invest the accumulated profit so far by closing the limits when they become profitable, while going into your direction. Those small counter trend corrections you see in ANY time frame I am referring to. In case the rate makes an u-turn, your profit is locked! There are more positives than negatives. I call this a brake pedal. Most, if not all EAs/systems, does not appear to have such. That is exactly why they high rev and then... one day... crash...
Maadik Hugiis. IQ 69.
  • #4
  • Feb 28, 2016 1:12pm Feb 28, 2016 1:12pm
Quoting feline207
Its just ridiculous. You hired 12 analysts and you analyzed the double tops and head and shoulders. The problem is you havent analyzed the 20/50 sma crossover. And also the rsi, you know - above 50 buy, below 50 sell. That should get you on the right path.
Sorry for my straightforwardness, but probably you didn't quite understand what I wanted to discuss. Head and shoulders / double top, the triangle are outdated models. Don't need to write about moving averages settings if all long time know that MA are well analyze only the history, but in the future are often mistaken as the RSI.
I write about what I'm trying to develop new models that won't be like all things we knew before. New models of tech analysis, their may be thousands and they are based on iteration. A bit later I'll post the charts with examples.
  • #5
  • Edited Feb 29, 2016 6:56am Feb 28, 2016 1:20pm | Edited Feb 29, 2016 6:56am
Quoting MoneyZilla
Matt, I think the right username you should use is Math, instead! No irony here. One way of making more pips is to start placing limit orders. The rate always bounce back enough. Even in a trend. These limit orders lock down your profit anyway and now you can invest the accumulated profit so far by closing the limits when they become profitable, while going into your direction. Those small counter trend corrections you see in ANY time frame I am referring to. In case the rate makes an u-turn, your profit is locked! There are more positives...
Hedging using limit orders is the same thing as closing in TP. I believe this trade doesn't solve basic task as to close the order at the highest point, and get the biggest profit.
  • #6
  • Feb 28, 2016 1:47pm Feb 28, 2016 1:47pm
Quoting feline207
Its just ridiculous. You hired 12 analysts and you analyzed the double tops and head and shoulders. The problem is you havent analyzed the 20/50 sma crossover. And also the rsi, you know - above 50 buy, below 50 sell. That should get you on the right path.
Hi matt, i dont really understand what your real problems are. is there in predicting in forex trading in which pairs, or in commodities, or in the way your are trading. i do make my very own indicators to support my trading needs. just for remember, as we know its almost impossible to buy at the lowest price and sell at the highest price. there is no holly grail to provide us for that moments. i do analyzed the market and always measure the market with the help of my indicators. and still sometimes i lose the trade. noone perfect.. i suggest you to evaluate your trading methods weekly to know n improving your trading skill. disiplined with our trading plan, and focusing in return of our investment. tq

best regard
Yustinus Saputra
  • #7
  • Feb 29, 2016 6:22am Feb 29, 2016 6:22am
Quoting Matt098
{quote} Hedging using limit orders is the same thing as closing in TP. I believe this trade doesn't solve basic task as to close the order at the highest point, and get the biggest profit.
Yes, it does. It gives you chance to lock down your profit and then close the limit orders when they become profitable and do this multiple times... Here is one guy (GoodBoy), doing only the limit orders. He is the Trading God Himself from the Sky. He did start with $4,000. He is adding some 100-150% every day. Like 4-6k a day. And that is exactly what he does with the limit orders. But he feels the pulse of the market, something I wish I had, but probably never will. I am just giving you that example, just to show you there is nothing to worry about. Obviously you do know what you are doing...
Maadik Hugiis. IQ 69.
  • #8
  • Feb 29, 2016 7:06am Feb 29, 2016 7:06am
Quoting Matt098
Hi traders and analysts. I've been investing and trading for a long time. But I always worry about when I cannot take all the movement entirety. I think you know how it happens. You've opened a sell position, closed your 100 pips, but the market went on. Greed appears and says immediately "Come on, let's make more money". Many traders have blown up accounts because of. Long time I was thinking how to take away all the market movement and not just part of it. My colleagues, the same traders as I argue that the market is repeated. Namely, it is composed...
I think you're really right. Market is repeated.
Interestingly, I never thought that way. I'm always skeptical at the fact it's possible in principle to find such a repetition and to understand that this is exactly what you are looking for. Good luck searching.
Subscribed.
  • #9
  • Mar 1, 2016 5:22am Mar 1, 2016 5:22am
Hello again.
Below is an example of coincidences in history. This is a very simple example, but on this topic I'll show coincidences that are repeat nearly same as previously was.
Also will be presented to your attention in the future more complicated structural formations.
Attached Images
  • #10
  • Mar 1, 2016 5:54am Mar 1, 2016 5:54am
Here is your previous screenshot of E/J with the 20 nearest price action (of E/J M30, L2 distance) and their extension in the future. The red and blue strong lines are respectively the mean and the median of the extensions in the future. As you can see the variance is very high. Despite in this example the bearishness of the forecast was correct it is not always the case. When wrong it can keep being wrong quite a long time. How do you expect to profit from this?
Attached Image (click to enlarge)
Click to Enlarge Name: EURJPY-M30.png Size: 32 KB
No greed. No fear. Just maths.
  • #11
  • Mar 1, 2016 6:14am Mar 1, 2016 6:14am
Quoting Matt098
Hello again. Below is an example of coincidences in history. This is a very simple example, but on this topic I'll show coincidences that are repeat nearly same as previously was. Also will be presented to your attention in the future more complicated structural formations. {image} {image}
ok sorry if i didnt get it at first, but still i dont. what is the "coincidence" here ? two waves of selling, different pairs and different times.
  • #12
  • Mar 1, 2016 6:36am Mar 1, 2016 6:36am
Quoting PipMeUp
Here is your previous screenshot of E/J with the 20 nearest price action (of E/J M30, L2 distance) and their extension in the future. The red and blue strong lines are respectively the mean and the median of the extensions in the future. As you can see the variance is very high. Despite in this example the bearishness of the forecast was correct it is not always the case. When wrong it can keep being wrong quite a long time. How do you expect to profit from this? {image}
Hello Pipmeup,

Yes, that looks like full of noise..
I know you love to model the prices. Have you ever thought about modeling the centroid of triangle?
Using ZigZag, the charts are full of triangles. Instead of using x and y coordinates, just by using x values (prices), the centroid becomes (A + B + C)/3, using the extremes of the zigzag.
Do you think there would be a merit in this? Could this make the model less noisy?
Thanks...
  • #13
  • Mar 1, 2016 6:59am Mar 1, 2016 6:59am
Quoting GoldTheHun
{quote} Hello Pipmeup, Yes, that looks like full of noise.. I know you love to model the prices. Have you ever thought about modeling the centroid of triangle? Using ZigZag, the charts are full of triangles. Instead of using x and y coordinates, just by using x values (prices), the centroid becomes (A + B + C)/3, using the extremes of the zigzag. Do you think there would be a merit in this? Could this make the model less noisy? Thanks...
Do you mind drawing a little sketch? I sounds like you need the next swing to draw the current triangle.
No greed. No fear. Just maths.
  • #14
  • Mar 1, 2016 7:12am Mar 1, 2016 7:12am
Quoting PipMeUp
{quote} Do you mind drawing a little sketch? I sounds like you need the next swing to draw the current triangle.
Sure. Here is my work of art drawing!!!!
I hope it is clear enough...
Attached Image (click to enlarge)
Click to Enlarge Name: ZZ.png Size: 11 KB
  • #15
  • Mar 1, 2016 7:18am Mar 1, 2016 7:18am
Quoting GoldTheHun
{quote} Sure. Here is my work of art drawing!!!! I hope it is clear enough... {image}
I did an analysis once, long time ago. On zigzags LN((B - C) / (A - B)) has normal distribution characteristics, but I couldn't get anywhere with that info...
Than I tried linear regression of the following form:

Input 1: LN((B - C) / (A - B))
Input 2: LN((C - D) / (B - C))
Output: LN((E - D) / (D - C))

This was fairly successful in predicting the height of the upcoming swing but I couldn't find any trading angle to exploit this info...
Hopefully you can come up with a better and more useful model...
  • #16
  • Edited 7:07am Mar 2, 2016 6:26am | Edited 7:07am
Quoting barracuda7
{quote} just for remember, as we know its almost impossible to buy at the lowest price and sell at the highest price. there is no holly grail to provide us for that moments. i do analyzed the market and always measure the market with the help of my indicators. and still sometimes i lose the trade. noone perfect.. i suggest...
Thanks for your opinion.
Market is a reflection of the psychological crowd actions. If to find match patterns in the past, we can use them in future. Patterns are not always 100% visually similar, admit the possibility of deviation up to 10%. Number of candles in pattern may also differ by 10%. Further if we compare them in the future and the past, they may be worked out 100%. Other words is possible to guess entry and exit almost to a point. So we can take out the whole movement, but sometimes happens pattern works on half or takes place the reversal.
My analysts want to eliminate all options where the models won't be fulfilled. So, achieve the target of 10 profitable trades out of 10.
  • #17
  • Edited 6:45am Mar 2, 2016 6:27am | Edited 6:45am
Attached Image
Attached Image


The charts show a small part of the movement (the model) repeated more than 16 years ago. Possible deviations in the graphical displaying up to 10%.
  • #18
  • Mar 2, 2016 6:31am Mar 2, 2016 6:31am
Quoting PipMeUp
Here is your previous screenshot of E/J with the 20 nearest price action (of E/J M30, L2 distance) and their extension in the future. The red and blue strong lines are respectively the mean and the median of the extensions in the future. As you can see the variance is very high. Despite in this example the bearishness of the forecast was correct it is not always the case. When wrong it can keep being wrong quite a long time. How do you expect to profit from this? {image}
Hi, PipMeUp. Could you describe the working principle of your indicator... Thx
  • #19
  • Mar 2, 2016 6:36am Mar 2, 2016 6:36am
Quoting feline207
{quote} ok sorry if i didnt get it at first, but still i dont. what is the "coincidence" here ? two waves of selling, different pairs and different times.
You're right. Pairs different and tf too. But sometimes movements are repeated, patterns are repeated and history itself.
  • #20
  • Mar 3, 2016 1:19am Mar 3, 2016 1:19am
Quoting Matt098
{quote} Hi, PipMeUp. Could you describe the working principle of your indicator... Thx
I take all the sequences of N (=40) bars and search for the K (=20) nearest past sequences to the current one (K nearest neighbors algo). The metric I use is simply the Euclidian distance. The search is made through a VP-tree that is built once. I only consider the results in the past (before the test sequence). I only use the current TF of the current instrument. I tried using the Euclidian distance of the returns with no improvements. As the results were quite deceptive I didn't investigate any further (like multi-TF or multi-instruments or other metrics). May you make me change my mind

I had already posted a screenshot of this indicator that may make you think a bit deeper:
http://www.forexfactory.com/showthre...67#post7932467 => there are a bullish and a bearish scenario!

(If only I knew I could have been hired (paid?) to do this )

EDIT: here another pic with error shades and evolution in time: http://www.forexfactory.com/showthre...45#post8559845
No greed. No fear. Just maths.
0 traders viewing now