EURUSD Statistical Data Mining Discussion

If you don't know where to start, start here! Don't be afraid to ask questions.

Moderator: moderators

User avatar
Techguy
rank: <50 posts
rank: <50 posts
Posts: 16
Joined: Tue Nov 25, 2008 12:45 am
Reputation: 5
Location: Washington State
Real name: Denton Hess
Gender: None specified
Contact:

EURUSD Statistical Data Mining Discussion

Postby Techguy » Thu Mar 28, 2019 7:00 pm

This is a TRO inspired, personal exercise.

Moderators - if this is somehow stepping on someones toes, let me know and I will drop it.

This is an invitation to receive a copy of a Excel SS with 61,000+ bars of H1 data. That is about 10 years worth of data. 2009.01.02 to 2019.03.27. Source Metaquotes, with the more recent data sourced by OANDA.

I cannot upload .xlsm or .xlsx files, plus currently it is 10.2Mb, so you will have to request it via PM.
I will have to share it via Dropbox in the future. So you will need Dropbox.

My intent is to data mine the *fertilizer* out of the data and draw a few tasty trading ideas out of it.

I will update the "Golden Copy" of the spreadsheet (SS) as time permits and as group input is received and performed.

If your idea is "Top Secret" I will be glad to keep it confidential, but personally I would love to hear it, and crunch it.

The first interesting thing I learned was that 79.7% of the time a one hour bar will have a high of 1.27 pips, on average, and 2.55 pips 69% of the time.
High-Open EURUSD 01_03_2009-03_27_2019.PNG
]High-Open EURUSD 01_03_2009-03_27_2019
High-Open EURUSD 01_03_2009-03_27_2019.PNG (44.18 KiB) Viewed 456 times


This bin out of 61688 bars shows that 1 pip is hit the most frequently, but 3 pips is usually achievable. The rub is that ~ 2 pips gets gobbled up in spread.

Useful Link to generate Histograms: https://youtu.be/Giewd9yH4q0

Your questions, comments, and input is requested. Please post public data mining ideas here in this thread.
Attachments
High-Open Bins 0-30 , 40, 50, 60, 70, 80, 100, 120 - 320 (x20).PNG
Bins 0 through 30 pips distribution
High-Open Bins 0-30 , 40, 50, 60, 70, 80, 100, 120 - 320 (x20).PNG (17.82 KiB) Viewed 456 times
"Everything can be taken from a man but one thing: the last of human freedoms - to choose one's attitude in any given set of circumstances, to choose one's own way." Viktor Frankl

Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.

User avatar
LeMercenaire
rank: 1000+ posts
rank: 1000+ posts
Posts: 2355
Joined: Tue Sep 27, 2016 12:11 am
Reputation: 1784
Gender: None specified

Re: EURUSD Statistical Data Mining Discussion

Postby LeMercenaire » Thu Mar 28, 2019 7:46 pm

Techguy wrote:This is a TRO inspired, personal exercise.

Moderators - if this is somehow stepping on someones toes, let me know and I will drop it.

This is an invitation to receive a copy of a Excel SS with 61,000+ bars of H1 data. That is about 10 years worth of data. 2009.01.02 to 2019.03.27. Source Metaquotes, with the more recent data sourced by OANDA.

I cannot upload .xlsm or .xlsx files, plus currently it is 10.2Mb, so you will have to request it via PM.
I will have to share it via Dropbox in the future. So you will need Dropbox.

My intent is to data mine the *fertilizer* out of the data and draw a few tasty trading ideas out of it.

I will update the "Golden Copy" of the spreadsheet (SS) as time permits and as group input is received and performed.

If your idea is "Top Secret" I will be glad to keep it confidential, but personally I would love to hear it, and crunch it.

The first interesting thing I learned was that 79.7% of the time a one hour bar will have a high of 1.27 pips, on average, and 2.55 pips 69% of the time.
High-Open EURUSD 01_03_2009-03_27_2019.PNG

This bin out of 61688 bars shows that 1 pip is hit the most frequently, but 3 pips is usually achievable. The rub is that ~ 2 pips gets gobbled up in spread.

Useful Link to generate Histograms: https://youtu.be/Giewd9yH4q0

Your questions, comments, and input is requested. Please post public data mining ideas here in this thread.



I don't see how it would be stepping on anyone's toes. If you are interested in and willing to crunch numbers like this, then more power to you.

User avatar
Techguy
rank: <50 posts
rank: <50 posts
Posts: 16
Joined: Tue Nov 25, 2008 12:45 am
Reputation: 5
Location: Washington State
Real name: Denton Hess
Gender: None specified
Contact:

Re: EURUSD Statistical Data Mining Discussion

Postby Techguy » Sat Mar 30, 2019 7:05 pm

Currently cataloging a list of ideas to research and crunch.

I did a keyword search on 5 minute bar frequency distributions I came across this video:
https://www.linnsoft.com/videos/time-day-session-highlow

His data set was 1975 days worth of 1 minute bars of the ES Mini S&P 500 Futures Contract. The data showed empirically that there was a statistically significant chance that the high would occur in the very first minute (7.6%) of the session, 18% of the time in the first 15 minutes, and 23% in the first 30 minutes. 5.9% of the time the high occurred during the last 3 minutes of the session. 4.6% of the time the high of the day was in the very last minute of trading.

This will be an interesting data crunch: Forex session vs day of week.
I'm speculating that Forex high and low of the day is variable due to the fact that Forex instruments are largely news driven.

One big take away was to develop an indicator that will mark the last high of the session, and the last low. Easy enough to do.

... more on this later...
"Everything can be taken from a man but one thing: the last of human freedoms - to choose one's attitude in any given set of circumstances, to choose one's own way." Viktor Frankl

jeffpeters
rank: <50 posts
rank: <50 posts
Posts: 1
Joined: Thu Nov 07, 2019 6:27 am
Reputation: 0
Gender: None specified

Re: EURUSD Statistical Data Mining Discussion

Postby jeffpeters » Thu Nov 07, 2019 6:41 am

Thank you!

User avatar
TheRumpledOne
rank: 10000+ posts
rank: 10000+ posts
Posts: 13652
Joined: Sun May 14, 2006 9:31 pm
Reputation: 1935
Location: Oregon
Real name: Avery T. Horton, Jr.
Gender: None specified
Contact:

Re: EURUSD Statistical Data Mining Discussion

Postby TheRumpledOne » Thu Nov 07, 2019 3:12 pm

Techguy wrote:Currently cataloging a list of ideas to research and crunch.

I did a keyword search on 5 minute bar frequency distributions I came across this video:
https://www.linnsoft.com/videos/time-day-session-highlow

His data set was 1975 days worth of 1 minute bars of the ES Mini S&P 500 Futures Contract. The data showed empirically that there was a statistically significant chance that the high would occur in the very first minute (7.6%) of the session, 18% of the time in the first 15 minutes, and 23% in the first 30 minutes. 5.9% of the time the high occurred during the last 3 minutes of the session. 4.6% of the time the high of the day was in the very last minute of trading.

This will be an interesting data crunch: Forex session vs day of week.
I'm speculating that Forex high and low of the day is variable due to the fact that Forex instruments are largely news driven.

One big take away was to develop an indicator that will mark the last high of the session, and the last low. Easy enough to do.

... more on this later...


I wrote a frequency distribution to show this.
IT'S NOT WHAT YOU TRADE, IT'S HOW YOU TRADE IT!

Please do NOT PM me with trading or coding questions, post them in a thread.

Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.


Return to “beginners forum”