why have you accepted the price-predictability assumption? | Page 6 | 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?

  • #101
  • Dec 15, 2013 12:17pm Dec 15, 2013 12:17pm
Quoting Turveyd
{quote} Yo dude, did I hear you type you've given up on trading ?? Didn't understand why, if so ??

Hi Turveyd, as simple as I cracked it. I just do a bit of bodyguard work now and advise my collegue
SEE LINE,TRADE LINE..PRICE HAS TO GO SOMEWHERE,,, PRICE WILL GO SOMEWHERE.
  • #102
  • Dec 15, 2013 1:13pm Dec 15, 2013 1:13pm
Quoting PipMeUp
{quote} From my data set: M1 candles on EUR/USD from 28/11/2001 04:10:00 to 13/12/2013 21:58:00 UP 1734571 41.90% DOWN 1729001 41.77% DOJI 675587 16.32% S&P is upside biaised yet, there are still as many up than down: M1 candles on S&P500 from 19/08/2008 09:31:00 to 13/12/2013 21:14:00 UP 433274 34.32% DOWN 423747 33.56 DOJI 405261 32.10 The marginal probability is still 50-50% ---------------------------------------------------------------------------- There is a very good point made my Turveyd about momentum. Market are heteroskedastic (the variance...
Yes, but the E/U is already a ranging market so dont expect more from it, if you look at the W1 TF you can clearly see those nice swings.So it is probably a 50-50 case, but do the same for AUD/JPY or CHF/JPY or GOLD and you will see that those markets are very biased.So that fact doesnt necessarly mean that all markets will be in all cases.But rather that the default aspect of every market is ranging.Because there is a sell and a paired buy order.When trends appear that only means that somebody put a larger position against the trend and let the trend ride up while he waits the pop out of the bubble and ready to take profit when the trend reverses.Or it may be because interest rates fluctuation or whatever.

The point is that the basic aspect is ranging, and the trend is the anomaly, not the range , so the fact that you have an edge in a trending market is logical because its different and varies from the equilibrium state of the market, while a ranging is only a larger volatility based consolidation period.

The breakout strategies are fine in my book, but i think you can make money in all types of markets if you segment them correctly and use proper tehniques to each type of market.But i would still prefer a nice edge, so a biased market/trending is by start the better one.

But to recognize a strong trend i think the best method is not to draw trendlines and other BS, because they are subjective crap, but rather to count the bars, which give us an objective meter of the trend's strength and direction
"There's a sucker born every minute" - P.T. Barnum
  • #103
  • Dec 15, 2013 1:25pm Dec 15, 2013 1:25pm
Quoting Turveyd
{quote} Now look at how many times M1 bars ( I have candles ) go up in row or down and compare that to H4, you'll get a lot more 10 bar 1 direction moves then H4 charts. And no Momentum gives you direction, there is then noise within that direction, ie pull backs in a up move which are buying op's.
They more or less follow a 1/2^n decay in both M1 and H4. Perfectly in agreement with a bernouilli trial
Inserted Code
EUR/USD - H4 since 28/11/2001
Length  Count   Percentage
1   4979   51.25
2   2413   24.84
3   1198   12.33
4    616   6.34
5    263   2.71
6    139   1.43
7     63   0.65
8     19   0.20
9     10   0.10
10     8   0.08
11     7   0.07
12     0   0.00
13     1   0.01


EUR/USD - M1 since 28/11/2001
Length  Count   Percentage
1       926479   51.52
2       452405   25.16
3       219155   12.19
4       106660   5.93
5        50106   2.79
6        23609   1.31
7        10927   0.61
8         4907   0.27
9         2249   0.13
10        1006   0.06
11         482   0.03
12         223   0.01
13          75   0.00
14          40   0.00
15          14   0.00
16           7   0.00
17           2   0.00
18           2   0.00
19           0   0.00
20           1   0.00
21           0   0.00
22           1   0.00
No greed. No fear. Just maths.
  • #104
  • Dec 15, 2013 1:34pm Dec 15, 2013 1:34pm
Quoting PipMeUp
{quote} They more or less follow a 1/2^n decay in both M1 and H4. Perfectly in agreement with a bernouilli trial EUR/USD - H4 since 28/11/2001 Length Count Percentage 1 4979 51.25 2 2413 24.84 3 1198 12.33 4 616 6.34 5 263 2.71 6 139 1.43 7 63 0.65 8 19 0.20 9 10 0.10 10 8 0.08 11 7 0.07 12 0 0.00 13 1 0.01 EUR/USD - M1 since 28/11/2001 Length Count Percentage 1 926479 51.52 2 452405 25.16 3 219155 12.19 4 106660 5.93 5 50106 2.79 6 23609 1.31 7 10927 0.61 8 4907 0.27 9 2249 0.13 10 1006 0.06 11 482 0.03 12 223 0.01 13 75 0.00 14 40 0.00 15 14 0.00...

Truely shocked at how close those 2 are together.
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #105
  • Dec 15, 2013 1:36pm Dec 15, 2013 1:36pm
Quoting Proximus
{quote} Yes, but the E/U is already a ranging market so dont expect more from it, if you look at the W1 TF you can clearly see those nice swings.So it is probably a 50-50 case, but do the same for AUD/JPY or CHF/JPY or GOLD and you will see that those markets are very biased.So that fact doesnt necessarly mean that all markets will be in all cases.But rather that the default aspect of every market is ranging.Because there is a sell and a paired buy order.When trends appear that only means that somebody put a larger position against the trend and...

I trade Range Primarily then Trend in that order, check out my method's chart.

120sma is the middle line, but the ranges are there, so aiming to enter between the middle and green dotted line with the direction, then that varies as to which range to use depending on markets strength.
Attached Image (click to enlarge)
Click to Enlarge Name: GA Perfect Example.jpg Size: 398 KB
Nothing to it, but to do it!!! Stick to the plan FOOL!!!!
  • #106
  • Dec 15, 2013 1:37pm Dec 15, 2013 1:37pm
Quoting Proximus
{quote} Yes, but the E/U is already a ranging market so dont expect more from it, if you look at the W1 TF you can clearly see those nice swings.So it is probably a 50-50 case, but do the same for AUD/JPY or CHF/JPY or GOLD and you will see that those markets are very biased.So that fact doesnt necessarly mean that all markets will be in all cases.But rather that the default aspect of every market is ranging.Because there is a sell and a paired buy order.When trends appear that only means that somebody put a larger position against the trend and...

On M1
--------
AUD/JPY
UP 1857284 43.310%
DOWN 1829505 42.663%
DOJI 601521 14.027%

CHF/JPY
UP 1827024 42.370%
DOWN 1809107 41.955%
DOJI 675894 15.675%

XAU/USD
UP 729773 47.514%
DOWN 725379 47.228%
DOJI 80755 5.258%



Weekly
--------
AUD/JPY
UP 345 54.936%
DOWN 281 44.745%
DOJI 2 0.318%

CHF/JPY
UP 350 55.732%
DOWN 278 44.268%
DOJI 0 0.000%
XAU/USD

UP 122 53.043%
DOWN 108 46.957%
DOJI 0 0.000%


Be careful at taking any serious conclusion about W1, there are only a handful of samples. Yet you are absolutely right that the fundamentals will always overcome the noise on the long term. This drift changes over time.
No greed. No fear. Just maths.
  • #107
  • Dec 15, 2013 1:38pm Dec 15, 2013 1:38pm
The non random part of this market is the voids in interest.if you locate the average price from monthly all the way down to 15 minutes and take a high extreme and low extreme .adjusting as time passes .these are your reference points .they consider multiple variables.

You can see as to whether the market is behaving randomly. These levels begin to converge.into 3 price zones.mean high and low.price will trade randomly within one of these zones.the unknown variables at this present moment are to us random. But the market has a view point from collective behavour pattern.which give us the 3 zones.these final variables are present this moment sentiment.this will eventually show its bias and quickly travel in a uniformed structured way to the upper or lower zone.the lack of open interest in this middle ground is lack of liquidity.this initself is inefficient market.
  • #108
  • Dec 15, 2013 1:58pm Dec 15, 2013 1:58pm
Here's a couple more price charts produced by randomly generated numbers, using Tradestation's RAND function. I did not write the code myself. I've looked over a lot of these random price charts and, as an experienced technical analyst, I can't distinguish between these charts and actual financial charts that I analyze, other than some of the bar sizes appear too small. All the technical patterns that I am familiar with and trade are present, however. This has blown my mind this weekend. See my earlier post in this thread if you haven't already.

In hindsight, these two charts look like they would be easy to trade. Just like a typical financial chart, however, not all of the price series trended so nicely. Just curious, for those of you better versed in statistics than I am, is it possible to develop technical strategies that when backtested over very large time period could be profitable in an inefficient market where all the technical variables are random?

Attached Image (click to enlarge)
Click to Enlarge Name: Random_Generated_Trend1.jpg Size: 76 KB


Attached Image (click to enlarge)
Click to Enlarge Name: Random_Generated_Trend2.jpg Size: 71 KB
  • #109
  • Dec 15, 2013 2:22pm Dec 15, 2013 2:22pm
Quoting PipMeUp
{quote} On M1 -------- AUD/JPY UP 1857284 43.310% DOWN 1829505 42.663% DOJI 601521 14.027% CHF/JPY UP 1827024 42.370% DOWN 1809107 41.955% DOJI 675894 15.675% XAU/USD UP 729773 47.514% DOWN 725379 47.228% DOJI 80755 5.258% Weekly -------- AUD/JPY UP 345 54.936% DOWN 281 44.745% DOJI 2 0.318% CHF/JPY UP 350 55.732% DOWN 278 44.268% DOJI 0 0.000% XAU/USD UP 122 53.043% DOWN 108 46.957% DOJI 0 0.000% Be careful at taking any serious conclusion about W1, there are only a handful of samples. Yet you are absolutely right that the fundamentals will always...
Or it may be because the absorbtion of the doji-like candles.As you go higher timeframe, you merge a few neighbor candles into 1, thus all doji's will be absorbed into a normal candle sooner or later.Like i told you in the PM, the candlestick distribution is like a binary tree, the all time forex history can be merged into 1 single candle, which is the root node, and as you go down, the smallest base, the leafs will be the tick data.And since the tick data is around 60% made up only by doji's as you go higher in timeframes the less and less dojis appear.So if you would have a 10 year timeframe you have a 99.999999999999999999% that that candle will not be a doji otherwise that would signal that absolute no market participation was made for 10 year which is very unlikely

But there is a generic bias regardless of doji's, dojis are only the sign of no market participation or low liquidity or low volatility, anyway you put it, but they are somewhat disrupting, and very annoying.

Any strategy i tried in the past, the statistics were always messed up because too many doji's appeared.So that is why higher timeframes appear to have better results because the absense of doji's.If the market had proper liquidity then all market timeframes would have an equal chance, i guess we need 5 more trillion dollars for that
"There's a sucker born every minute" - P.T. Barnum
  • #110
  • Dec 15, 2013 2:47pm Dec 15, 2013 2:47pm
Quoting Proximus
{quote} Or it may be because the absorption of the doji-like candles.As you go higher time frame, you merge a few neighbour candles into 1, thus all doji's will be absorbed into a normal candle sooner or later.
Well, I'm not quite sure about this assumption about doji's necessarily merging into a standard candle. Looking back over weekly and monthly time frames for a few pairs, I've seen combinations of standard candle shapes actually combine into a classic doji candle. Therefore you could say indecision or changes of balance during that time period caused the 're-balancing' of the upward and downward candles.

We also mustn't forget that we calculate candles based on specific slices of how we measure time. If we create alternate candles that were offset by 10%, either earlier or later, then that will tweak the shape of our candles...

Interesting ideas though. I will admit that only follow some of what's been stated within this thread, regardless, I do appreciate people taking the time and effort to post their thoughts along these lines. I'll keep reading and see what I can pick up.
  • #111
  • Dec 15, 2013 11:42pm Dec 15, 2013 11:42pm
Quoting Nefser
{quote} Well, I'm not quite sure about this assumption about doji's necessarily merging into a standard candle...
They do, if by doji you mean OPEN[n] =CLOSE[n], if you talk about doji-like bars like spinning top and others, then it probably it wont be that high impact since ranging periods often result in a spinning top in higher timeframe, but the doji is a low liquidity period, which gets eliminated as you go to higher timeframe.

To explain it better, here is a part of my doji research done on a 10 year span, you can clearly see what i meant

Attached File(s)
File Type: xlsx DOJI RESEARCH.xlsx   26 KB | 364 downloads
"There's a sucker born every minute" - P.T. Barnum
  • #112
  • Dec 16, 2013 3:25am Dec 16, 2013 3:25am
I thought about this doji thingy a little bit more. If I flip 2 coins what is the probability I get as many heads than tails?
There are 4 possibilities {HH, HT, TH, TT}, two of them have as many heads than tails {HT, TH}. The probability is 2 out of 4; 50%.
If I flip 4 coins? 6, 8, 10...? The probability is p=C(n, n/2) / 2^n where C(n, k) is the binomial coefficient.

The more coins, the less dojis...

Inserted Code
Coins  heads=tails      total   percentage
  2              2           4  50.000%
  4              6          16  37.500%
  6             20          64  31.250%
  8             68         256  26.562%
 10            252        1024  24.609%
 12            924        4096  22.558%
 14           3432       16384  20.947%
 16          12864       65536  19.628%
 18          48590      262144  18.535%
 20         184756     1048576  17.619%
 22         703302     4194304  16.768%
 24        2704142    16777216  16.117%
 26       10400598    67108864  15.498%
 28       40116532   268435456  14.944%
 30      155117490  1073741824  14.446%
Attached Image
No greed. No fear. Just maths.
  • #113
  • Dec 16, 2013 5:21am Dec 16, 2013 5:21am
Quoting PipMeUp
I thought about this doji thingy a little bit more. If I flip 2 coins what is the probability I get as many heads than tails? There are 4 possibilities {HH, HT, TH, TT}, two of them have as many heads than tails {HT, TH}. The probability is 2 out of 4; 50%. If I flip 4 coins? 6, 8, 10...? The probability is p=C(n, n/2) / 2^n where C(n, k) is the binomial coefficient. The more coins, the less dojis... Coins heads=tails total percentage 2 2 4 50.000% 4 6 16 37.500% 6 20 64 31.250% 8 68 256...

Could be but the timeframes in MT4 are not aligned like that.That flip in MT4 would correspond to this:

M1= 2 coin
M5= 10 coin
M15 = 30 coin
M30 = 60 coin
H1 = 120 coin

So that being said it would look like this

Attached Image (click to enlarge)
Click to Enlarge Name: a.png Size: 12 KB


It is almost like my research, with a correlation almost identical.I think this is not a coincidence.

So to sum up, this would suggest that doji's form 100% randomly
"There's a sucker born every minute" - P.T. Barnum
  • #114
  • Dec 17, 2013 8:16am Dec 17, 2013 8:16am
Quoting the redlion
{quote} I will once I learn how to lol, then I will post it here. would that be a fair way to describe a Cauchy Distribution? http://en.wikipedia.org/wiki/Cauchy_distribution
H4 pips return for EUR/USD since 28/11/2001 against Cauchy.
Attached Image (click to enlarge)
Click to Enlarge Name: qqplotCauchy.png Size: 9 KB
No greed. No fear. Just maths.
  • #115
  • Dec 18, 2013 7:05am Dec 18, 2013 7:05am
Quoting nubcake
{quote} I dunno. Test me again. And if you recall I said to you, honestly (if you believe me or not), that whatever recent chart you had decided to pick as your true chart would not be something I had seen because I had simply not been looking at any charts at all for quite sometime. I was not familiar with any recent, or semi-recent movements or anything. Whether anyone believes that or not is up to them. There is a certain swagger and timing that the market tends to make, at least as far as I'm concerned, that 'random' charts don't tend to make...
May I ask you to try with these three charts please?
Attached Image (click to enlarge)
Click to Enlarge Name: TestChart-1.png Size: 15 KB
Attached Image (click to enlarge)
Click to Enlarge Name: TestChart-2.png Size: 14 KB
Attached Image (click to enlarge)
Click to Enlarge Name: TestChart-3.png Size: 14 KB
No greed. No fear. Just maths.
  • #116
  • Dec 18, 2013 7:45am Dec 18, 2013 7:45am
Sorry, actually i think the 3 charts are real.
  • #117
  • Dec 18, 2013 10:51am Dec 18, 2013 10:51am
Quoting PipMeUp
{quote} May I ask you to try with these three charts please? {image}{image}{image}
number 3? perhaps, but let's see if nubcacke gets it right
AVT INVENIAM VIAM AVT FACIAM
  • #118
  • Dec 18, 2013 11:10am Dec 18, 2013 11:10am
Quoting PipMeUp
{quote} May I ask you to try with these three charts please? {image}{image}{image}
my take, 1 and 3.
Trading = a mirror to your human flaws. Fix them or be fixed.
  • #119
  • Dec 18, 2013 8:00pm Dec 18, 2013 8:00pm
All of them are real.
  • #120
  • Dec 18, 2013 8:18pm Dec 18, 2013 8:18pm
Quoting Hairball
Here's a couple more price charts produced by randomly generated numbers, using Tradestation's RAND function. I did not write the code myself. I've looked over a lot of these random price charts and, as an experienced technical analyst, I can't distinguish between these charts and actual financial charts that I analyze, other than some of the bar sizes appear too small. All the technical patterns that I am familiar with and trade are present, however. This has blown my mind this weekend. See my earlier post in this thread if you haven't already....

Sorry but i cant understand how can one generate random numbers using algorithms or code ?
0 traders viewing now