Statistical Probability Trading With Price Action | Page 21 | Forex Factory

Options

Go to
Permalinks
Print Thread

Related Instruments

GBP/USD

Similar Threads

GbpUsd Statistical Probability With Price Action 109 replies

EurJpy Statistical Probability Trading With Price Action 46 replies

GbpJpy Statistical Probability With Price Action 22 replies

Statistical Validation on trading 24 replies

Attachments: Statistical Probability Trading With Price Action
Exit Attachments

Statistical Probability Trading With Price Action

  • #401
  • Jul 29, 2007 8:13pm Jul 29, 2007 8:13pm
Possible scenario of gbpusd hitting between 2.0365 and 2.0390 based on the present low 2.0181, weekly retracement probabilites, weekly range of motion probabilites, and 5 day S&R in early July. This may be a mid week target with possible more down side movement there after. Closed out last short position at 2.0214. Wait to see price action when the London session opens.
Attached Image
  • #402
  • Jul 29, 2007 9:10pm Jul 29, 2007 9:10pm
Possible scenario the gbpjpy based on fib retracements and weekly range of motion. Mid week scenario then most are saying more down. Just closed with a 4hr pin.
Attached Image
  • #403
  • Jul 29, 2007 9:14pm Jul 29, 2007 9:14pm
Revised scenario after establishing a present low of the day of 118.01
Attached Image
  • #404
  • Jul 29, 2007 9:20pm Jul 29, 2007 9:20pm
Possible scenario. Closed with a 4hr bull pin bar. Looking for price action at the London open.
Attached Image
  • #405
  • Jul 29, 2007 9:41pm Jul 29, 2007 9:41pm
So far there has only been a 22 pip retracement into last weeks weekly bar. If you look at the weekly bar to the right of the previous weeks bar there is always price retracement or price action in between the high and low of the previous week. So far we have only about 22 pips of price action in last weeks bar. Looking to see more retracement into last weeks bar before more down side? Note price the follow week of Feb 25's big bear run.
Attached Image
  • #406
  • Jul 29, 2007 9:56pm Jul 29, 2007 9:56pm
Scenario two with it bouncing off around the 23.6 weekly fib which has a 96% probability of being hit and the early July 5 day S&R levels then moving south. See post 428 for scenario 1.
Attached Image
  • #407
  • Jul 30, 2007 3:08am Jul 30, 2007 3:08am
Not saying it is fully extended but look at these retracement levels and probabilities at the top of the daily run. These are daily statistical probability fib retracements.
Attached Image
  • #408
  • Jul 30, 2007 3:11am Jul 30, 2007 3:11am
Daily Range for GbpUsd for the year
Attached Image (click to enlarge)
Click to Enlarge Name: GbpUsd Daily Range.jpg Size: 28 KB
  • #409
  • Jul 30, 2007 3:20am Jul 30, 2007 3:20am
Daily Range For GbpJpy
Attached Image (click to enlarge)
Click to Enlarge Name: GbpJpy Daily Range.jpg Size: 21 KB
  • #410
  • Jul 30, 2007 3:34am Jul 30, 2007 3:34am
GbpChf this week possible scenario
Attached Image
  • #411
  • Edited 6:11am Jul 30, 2007 4:07am | Edited 6:11am
Quoting goodthings
The following are S&R statistics that can be used as a guide of when to exit trades based on the daily range average
Since Jan of 2007 GbpUsd touches with in 10 pips of the S1 or R1 82% of the time the EurJpy 87% of the time
Since Jan of 2007 GbpUsd touches with in 10 pips of the S2 or R2 33 % of the time the EurJpy 39% of the time
Since Jan of 2007 GbpUsd touches with in 10 pips of the S3 or R3 20 % of the time the EurJpy 19% of the time
I believe more reliable statistical S and R would be statistical S/R - during (for example) last year price reached above 50% of days and below 50% of days - this level indicates real S/R when price bounced 50% of cases, average daily range imho means less.

To calculate this (example R1) we should calculate High - Open on serie of days ( lets say 300 days) and then count number of 1, 2, 3, 4,5, etc (histogram) - number of occurences of 1 pip, 2 pip above close. Lets say we denote it as O(1), O(2) etc.

Then we should divide number of days in two, so we have 150.

Then start adding number of occurences O(1) + O(2) + O(3) + O(4) + ....

When the sum reaches 150 (Lets say it would be at 43 pips) we have statistically significant resistance level.

At this level over our sample of 300 days - 150 times price got lover HI of day and 150 times higher.

At this level we have 50% chance that this will be height of the day. At higher levels - lets say 70% we have 70% chance that this is the highest point of the day.

Here percentage is a more or less good estimation of probability over a series of data.

I suppose that R calculated from average daily range will be greater than statistical.

Example which is better - suppose that during a week we have

1 day - 30
2 day - 30
3 day - 30
4 day - 30
5 day (non farm paytroll) 100 pips.

Statistical R would be 30.

Based on average daily range (4x30+100)/5 = 220/5 = 44

Hope you get the idea

P.S (maybe instead of high - open of the current day better will be high minus close of the previous day.)
Attached File(s)
File Type: zip histogram USD JPY 7 years.zip   19 KB | 436 downloads
Quidquid latine dictum sit altum viditur
  • #412
  • Jul 30, 2007 7:34pm Jul 30, 2007 7:34pm
Daily Statistical Probabilities if current low 241.34 holds. Looking to see a break above the current daily high on the 4hr and the 23.6 fib.
Attached Image
  • #413
  • Jul 30, 2007 7:42pm Jul 30, 2007 7:42pm
Looking at 2.031 the daily 80 pips with 70% statistical probability of being hit and also the weekly 200 pips with a 77% statistical probability of being hit. Just above the weekly fib of 23.6 which has a 96% probability of being hit. Next resistance level the 2.0365 the five day s/r in early mid July. Then a possible mid week retracement Wed or so.
Attached Image
  • #414
  • Jul 30, 2007 7:56pm Jul 30, 2007 7:56pm
Notice how many times the GbpUsd had a mid week pivot on Tuesday or Wendesday. Only twice did the weekly run all week to Friday.
Attached Image
  • #415
  • Jul 30, 2007 8:44pm Jul 30, 2007 8:44pm
Not ignoring your post give me a few days to review it and comment. Thanks

Tom

Quoting dottore
I believe more reliable statistical S and R would be statistical S/R - during (for example) last year price reached above 50% of days and below 50% of days - this level indicates real S/R when price bounced 50% of cases, average daily range imho means less.

To calculate this (example R1) we should calculate High - Open on serie of days ( lets say 300 days) and then count number of 1, 2, 3, 4,5, etc (histogram) - number of occurences of 1 pip, 2 pip above close. Lets say we denote it as O(1), O(2) etc.

Then we should divide number of days in two, so we have 150.

Then start adding number of occurences O(1) + O(2) + O(3) + O(4) + ....

When the sum reaches 150 (Lets say it would be at 43 pips) we have statistically significant resistance level.

At this level over our sample of 300 days - 150 times price got lover HI of day and 150 times higher.

At this level we have 50% chance that this will be height of the day. At higher levels - lets say 70% we have 70% chance that this is the highest point of the day.

Here percentage is a more or less good estimation of probability over a series of data.

I suppose that R calculated from average daily range will be greater than statistical.

Example which is better - suppose that during a week we have

1 day - 30
2 day - 30
3 day - 30
4 day - 30
5 day (non farm paytroll) 100 pips.

Statistical R would be 30.

Based on average daily range (4x30+100)/5 = 220/5 = 44

Hope you get the idea

P.S (maybe instead of high - open of the current day better will be high minus close of the previous day.)
  • #416
  • Jul 30, 2007 11:00pm Jul 30, 2007 11:00pm
Area of S/R in mid July
Attached Image
  • #417
  • Edited 3:00pm Jul 31, 2007 2:49am | Edited 3:00pm
Quoting goodthings
Not ignoring your post give me a few days to review it and comment. Thanks

Tom
Hi Tom, I changed my post several times due to some errors and I'm not sure if I corrected all, so please if something is unclear ask, and dont waste your valuable time to comment my errors. I see how excellent service you give to community and I'm impressed.

If you want to comment please comment the general idea: I believe that over a sample of days (any timeframe) we can calculate probability in this way:

P(x) = P(price reaches not far than x) = (number of days that price reached x or less) divided by (number of days in the sample).

So I counted how many times price reached 1 pips above open, 2 pips above open, 3 pips above open, .... 200 pips above open (there are mine O(1), O(2) etc) and then:

my sample of days was 2108.

Then I started adding -
number of days price reached 1 pips above open,
plus number of days price reached 2 pips above open,
plus number of days price reached 3 pips above open,
plus number of days price reached 4 pips above open, ...
.....
plus number of days price reached n pips above open.

This is SUM (n) = O(1) + O(2) + O(3) + O(4) +... O(n)

at n = 37 I got 1057 which is close to half of 2108

It means that Hi of the day was below this number (37) on half of days, and on half of days price (HI of the day) was above.

Using this system I can calculate probability (lets say this is an estimation of probability) that price reaches certain level - I can give levels of 10, 20, 30 ... 60, 70, 80, 90 probability.

Thanks for your time.
Quidquid latine dictum sit altum viditur
  • #418
  • Edited 7:04pm Jul 31, 2007 5:25pm | Edited 7:04pm
The UsdJpy has bounced off the 38.2 weekly fib of last week and so far has ranged within the weekly 70% weekly statistical probability of 150 pips. Most currencies are ranging beyond the 70% normal range. Usd weakness could bring further down side for the week. Grey area is today's daily range. See post 422 and 427 for earlier analysis
http://www.forexfactory.com/showthre...=27624&page=29
Attached Image
  • #419
  • Jul 31, 2007 5:48pm Jul 31, 2007 5:48pm
Came with in 14 pips of the weekly fib of 38.6 (there is a 10 pip tolerance to call it a hit). Also it is the 50% statistical probability (SB) of the weekly range of 260 pips 89 ema and the 5day s/r for early July. Went 23 pips beyond the daily 51% (SB) range of motion and retraced beyond the daily 61.8 fib. The gray shaded area is this past trading day price action on the 4hr chart. See the following link for earlier posts for the GbpUsd. http://www.forexfactory.com/showthre...=27624&page=28.
Attached Image
  • #420
  • Jul 31, 2007 6:00pm Jul 31, 2007 6:00pm
Dollar week more up side and beyond the (50%) 260 pip weekly SB of range of motion. With the market motion things see to be ranging broader then normal. Then possilbly mid week down side or tight range of motion to finish out the week. More up side would be nice I am long gbpusd
Attached Image
0 traders viewing now