why have you accepted the price-predictability assumption? | Page 23 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

So I accepted a Prop Trading job in South Beach Miami 44 replies

Why you shouldn't show how much you have made in FX 35 replies

Incorporating Slippage Assumption in backtesting 4 replies

Market Predictability - Technical Signals 25 replies

Contesting the Accepted View of Bernanke 0 replies

Attachments: why have you accepted the price-predictability assumption?
Exit Attachments

why have you accepted the price-predictability assumption?

  • #441
  • Dec 26, 2013 9:49pm Dec 26, 2013 9:49pm
Quoting Magix
{quote} If I could show you the way to disprove your randomness, would you take the time to learn?
Well I for one certainly would

I would like to see how one can "predict" price. I am aware of all the candle formations and nearly every indicator available for retail consumption.

Plus it would make a damned good discussion! Building millipede takes a lot of time, and I would love to learn a faster way that is just as safe.
If you are good at something, never do it for free--Joker
  • #442
  • Dec 26, 2013 10:35pm Dec 26, 2013 10:35pm
Quoting jmn5611
would love to learn a faster way that is just as safe.

Thing is, knowing where price will tend, doesn't make it safe, it is only a single piece of information that if used incorrectly, will still drain your account.

Have you heard of somebody,or have you, yourself, blown up an account playing the trend?

I have on both. All while picking with absolute pinnacle precision, I have personally lost 2 accounts. I have also watched and coached others who have done the same at one point in their trading careers.

Take these targets, couple them with auction market, supply and demand, cross and multiple pair,time frame and session analysis and you can be fluid and in tune with the pulse of price.

Targets...razor the wicks off your hourly and weekly candles...you should be able to see them from there.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #443
  • Dec 27, 2013 12:09am Dec 27, 2013 12:09am
Quoting Magix
Targets...razor the wicks off your hourly and weekly candles...you should be able to see them from there.
Here is a sample of "Random", vs Actual predictability.

Or, what you see on a naked chart, and what I see using 2 time frames and candle bodies.

2 breakzones, 2 price fills.

2 places of price predictability.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: uj1.png Size: 38 KB
Click to Enlarge Name: uj2.png Size: 46 KB
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #444
  • Dec 27, 2013 12:28am Dec 27, 2013 12:28am
Quoting Magix
{quote} Thing is, knowing where price will tend, doesn't make it safe, it is only a single piece of information that if used incorrectly, will still drain your account. Have you heard of somebody,or have you, yourself, blown up an account playing the trend? I have on both. All while picking with absolute pinnacle precision, I have personally lost 2 accounts. I have also watched and coached others who have done the same at one point in their trading careers. Take these targets, couple them with auction market, supply and demand, cross and multiple...
Position sizing has caused me to damage some accounts very badly. I'm with Oanda, so technically you can't "blow up". This is why I now just go in the direction price went yesterday, and last week. As far as picking direction, sure, I'm usually right, eventually. Eventually can cause more draw down than can be withstood however, so I have taken to scaling into a trend as it emerges.
If you are good at something, never do it for free--Joker
  • #445
  • Dec 27, 2013 12:31am Dec 27, 2013 12:31am
Quoting Magix
{quote} Here is a sample of "Random", vs Actual predictability. Or, what you see on a naked chart, and what I see using 2 time frames and candle bodies. 2 breakzones, 2 price fills. 2 places of price predictability. {image} {image}
That a pretty good idea to plot H1 candles over M5 candles. What is the indy for that?
If you are good at something, never do it for free--Joker
  • #446
  • Dec 27, 2013 12:37am Dec 27, 2013 12:37am
Quoting jmn5611
{quote} That a pretty good idea to plot H1 candles over M5 candles. What is the indy for that?
Indicator called Large Time Frame, find it posted on the link HERE.

The template is just observation of the DTEMA.

6 pairs. EURGBP, USDJPY, EURUSD, EURJPY, GBPUSD, GBPJPY.

EURGBP and USDJPY <-Use these for speed of trade on the other 4 pairs.

Watch for supply and demand zones on the 4 respective and play the strongest in favor of the trend.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #447
  • Dec 27, 2013 12:43am Dec 27, 2013 12:43am
Quoting jmn5611
{quote} Position sizing has caused me to damage some accounts very badly. I'm with Oanda, so technically you can't "blow up". This is why I now just go in the direction price went yesterday, and last week. As far as picking direction, sure, I'm usually right, eventually. Eventually can cause more draw down than can be withstood however, so I have taken to scaling into a trend as it emerges.
That's why I chip the smaller ranges with a higher risk.

If I am wrong, and it does happen, I can easily close out for a small loss and watch or pick a different set.

Here is another example. The BZ at the bottom is a weekly. The one at the top is to the M5.

If I was to want to scale in, that might be a place to target, but as I am really selective about my seasonal trades, I'm kewl just sitting this shit out for now.
Attached Image (click to enlarge)
Click to Enlarge Name: EJ1.png Size: 37 KB
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #448
  • Dec 27, 2013 12:51am Dec 27, 2013 12:51am
UJ, Hammering down, hourly close below and the next demand range is in sight.

Of course, a trend confirmation will need be present...but again, it's more of a time thing than is price...and still needs that hourly close below support.
Attached Image (click to enlarge)
Click to Enlarge Name: uj3.png Size: 43 KB
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #449
  • Edited 1:26am Dec 27, 2013 1:15am | Edited 1:26am
Quoting Proximus
{quote} Oh boy, you are now confusing 2 different termonologies. RANDOMNESS DOESNT EQUAL NON-PREDICTABILITY ITS ONLY EQUAL IF ITS A 50-50% ...

but i never ever said that the market cannot be predicted just because its random
Good grief dude. I realize you are pushing some band wagon, but still.. such a bold (literally) absurd statement.

Try this.
Lottery drawings are random. Can you predict the next drawing? even 50% of the time as you are stating?

Also: "the market cannot be predicted just because its random"

That makes no sense.

If the market is random (which it is, most of the time). It cannot be predicted. period.
  • #450
  • Dec 27, 2013 2:00am Dec 27, 2013 2:00am
Quoting Mingary
{quote} Good grief dude. I realize you are pushing some band wagon, but still.. such a bold (literally) absurd statement. Try this. Lottery drawings are random. Can you predict the next drawing? even 50% of the time as you are stating? Also: "the market cannot be predicted just because its random" That makes no sense. If the market is random (which it is, most of the time). It cannot be predicted. period.
I think the market is ultimately random in that we cannot know if the trade we put on will work. But, what this thread is evolving into is showing that very educated guesses can be made.

If you see price hit a longer TF S/R and it bounces from the shorter TF then you have a low risk entry based on what we THINK other traders MIGHT do. If there are no bigger fish in the sea, then we will get our way.

So is the market random? Yes......

Can we make educated guesses that turn out right more often than not.... Yes.

Does that mean that the market is not random? No.....

Just some more stuff to chew on.....I am not the authority in any way....I trade as if I am flipping a weighted coin. Bias might change, but I still have not the slightest clue if my trade will work. I can only follow price and see where it takes me. That's all the skill I have.
If you are good at something, never do it for free--Joker
  • #451
  • Dec 27, 2013 2:03am Dec 27, 2013 2:03am
Quoting Magix
UJ, Hammering down, hourly close below and the next demand range is in sight. Of course, a trend confirmation will need be present...but again, it's more of a time thing than is price...and still needs that hourly close below support. {image}
This is good stuff Magix, thanks for sharing. It seems that you are dealing in probabilities on an instinct level. Have you tried to quantify your set ups?
If you are good at something, never do it for free--Joker
  • #452
  • Dec 27, 2013 2:42am Dec 27, 2013 2:42am
Quoting Magix
UJ, Hammering down, hourly close below and the next demand range is in sight. Of course, a trend confirmation will need be present...but again, it's more of a time thing than is price...and still needs that hourly close below support. {image}

Dude, we can all draw lines on after the event with 100% accuracy, how about calling some live ?
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #453
  • Dec 27, 2013 3:16am Dec 27, 2013 3:16am
proximus....I am truly perplexed on why you are taking so much heat, I thank you for your efforts and contribution to this thread. I will test your data for randomness, then I will test and compare it to the market.


Random coin flips when charted can produce the illusion of trends, and chart patterns, s/d, s/r zones, and can even bounce off moving averages and BB's, I am not sure what is making magix so upset.
AVT INVENIAM VIAM AVT FACIAM
  • #454
  • Dec 27, 2013 5:43am Dec 27, 2013 5:43am
Quoting the redlion
proximus....I am truly perplexed on why you are taking so much heat
Because...

Attached Image
  • #455
  • Dec 27, 2013 6:06am Dec 27, 2013 6:06am
I'm a little slow and would appreciate a some help: what exactly does the random number-generated chart model? What point(s) is (are) being made with it? I agree that random events impact market structure and Price Action. Not sure it answers any of the questions that red asked in his first post.
  • #456
  • Dec 27, 2013 7:25am Dec 27, 2013 7:25am
All this debate about random vs pseudo-random is non sense. This is a non-issue. We are using a few millions bits of data. That's a ridiculously small sample. Proximus results are bad because he is using a uniform distribution. Full stop. Let's move forward.

  1. Yes random and uniform can be put in the same sentence.
  2. Yes a random walk can go anywhere and will generate trends.
  3. Yes the ratio of heads and tails for a fair coin converges towards 1. But the difference between the number of heads and tails can be arbitrarily big. This number follows a random walk.
  4. Yes a biased coin will also generate a random walk (with drift)
  5. Yes a random event can be predicted, within a margin of error, provided you know its probability distribution and this pdf has a finite variance.
  6. Yes atmospheric data are random. Despite this fact specialists can forecast the weather.
  7. We can only approximate the distribution of natural events. They very often follow a Normal distribution. How tall are you? Your height almost follows a Normal distribution. But not exactly otherwise we should meet a few 5m tall people...
  8. The Normal distribution has a mean and a variance defined. As well as all the other moments.
  9. Cauchy distribution has no mean and infinite variance.
  10. Market aren't Normally distribution. They look much more Cauchy but they aren't Cauchy. See http://www.forexfactory.com/showthre...55#post7161355
  11. The variance of the market varies over time. See http://www.forexfactory.com/showthre...66#post7166266. Therefore the probability distribution also varies over the time.

Does the market have a finite variance?

Is the market autoregressive?

If it is autoregressive we know from post 27 that the coefficients are smaller than the noise level. Can we estimate them nonetheless?

Can we estimate the local probability distribution given the recent price action to provide us more information than the marginal probability distribution?

No greed. No fear. Just maths.
  • #457
  • Dec 27, 2013 7:26am Dec 27, 2013 7:26am
Quoting GEfx
I'm a little slow and would appreciate a some help: what exactly does the random number-generated chart model? What point(s) is (are) being made with it? I agree that random events impact market structure and Price Action. Not sure it answers any of the questions that red asked in his first post.
Some time ago, redlion randomly generated some charts and nubcake was able to distinguish which ones were real/random. If you have a look at proximus result, you immediately see they are fake. This is because the uniform distribution of the movements is irrealistic. In this model a doji has the same probability as a 150 pip bar.

Redlion used the Normal distribution. It is slighly nearer to the market movement, but the very big moves are not enough present. I tought nubcake was able to see the difference because these sudden big moves, long wicks or spikes, were absent from redlion's chart. I created random charts using the market itself. I though this could be enough to fool nubcake. To build the random charts each tick is drawn independently of the other ones from the same distribution I estimated from the data of E/U. The ticks are independent and identically distributed (iid).


Nubcake could distinguish a genuine chart from two random ones. He explained this was because he couldn't see/feel the "rythm" of the market in the random charts. This is a hint that the markets are certainly not independent. In layman terms, the past price action has an influence on the future of the market. How far in the future is the whole question.
No greed. No fear. Just maths.
  • #458
  • Dec 27, 2013 7:53am Dec 27, 2013 7:53am
Quoting PipMeUp
{quote} Some time ago, redlion randomly generated some charts and nubcake was able to distinguish which ones were real/random. If you have a look at proximus result, you immediately see they are fake. This is because the uniform distribution of the movements is irrealistic. In this model a doji has the same probability as a 150 pip bar. Redlion used the Normal distribution. It is slighly nearer to the market movement, but the...
Thanks PipMeUp, I caught all of that, but I thought I'd missed something. My conclusion is that this discussion is not very relevant to trading. After all, you would never trade a chart unless the underlying narrative of market structure and PA were clear. Such a chart would have no connection to the world, as the world is unfolding (the chart would zig when the world zagged).
  • #459
  • Dec 27, 2013 9:44am Dec 27, 2013 9:44am
Quoting GEfx
{quote} Thanks PipMeUp, I caught all of that, but I thought I'd missed something. My conclusion is that this discussion is not very relevant to trading. After all, you would never trade a chart unless the underlying narrative of market structure and PA were clear. Such a chart would have no connection to the world, as the world is unfolding (the chart would zig when the world zagged).
This was exactly my point in that there really is, no point.

Without knowing trade pair, sessions when the data gets printed and even down to the day of the week, they are irrelevant.

An fruitless quest.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #460
  • Dec 27, 2013 9:49am Dec 27, 2013 9:49am
Quoting jmn5611
{quote} This is good stuff Magix, thanks for sharing. It seems that you are dealing in probabilities on an instinct level. Have you tried to quantify your set ups?
I am not exactly sure what you mean by quantifying my set ups...

My trading is based on more of an IF-Then.

These just offer areas to take trades based on previous market relevant areas.

Kind of the Chart Whispers.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
2 traders viewing now