Sunday, January 31, 2016

Analysis of DAX Initial Balance

In a similar manner to my analysis of Initial Balance on FTSE100 (here), I've had a look at the DAX.

Total number of days analysed: 1010
Average High-Low range: 132.3, standard deviation 75.5
Average IB range: 56.9, standard deviation 33.9
IB High (IBH) was broken: 57%
IB High 50% extension (IBHX1) was broken: 36%
IB High 100% extension (IBHX2) was broken: 21%
IB High 200% extension (IBHX3) was broken: 7%
IB Low (IBL) was broken: 58%
IB Low 50% extension (IBLX1) was broken: 39%
IB Low 100% extension (IBLX2) was broken: 24%
IB Low 200% extension (IBLX3) was broken: 10%

Note 1 : Data obtained from modified SessionLines indicator
Note 2 : Analysed using Excel
Note 3 : Test for a "break" allowed for 10 pips leeway, e.g. had to break the IBH by 10 pips to be counted

Looking in more detail at the ratio of High-Low range to IB range, I generated the following data.



The % column is the proportion of days that fall into a particular "bin", e.g. for the bin labelled 1.25 it counts those days where the R/IB ratio is between 1.00 and 1.25. The data is left-skew, with a peak at around 2.0 (i.e. 200%), but also quite high proportion up to an R/IB ratio of 3. Days with R/IB ratio of 1 (i.e. the IB range was approx 100% of the total range) were very few.

This type of data gives a feel for what might be expected in terms of normal ranges, based off the 1st hour of trading.

On days, when the IBL was not broken (by more than 10 pips), the following stats are obtained...

IB High (IBH) was broken: 87%
IB High 50% extension (IBHX1) was broken: 62%
IB High 100% extension (IBHX2) was broken: 38%
IB High 200% extension (IBHX3) was broken: 13%
This happens on about 40% of all days analysed.


Similarly, on days when the IBH was not broken (by more than 10 pips), the following stats are obtained...

IB Low (IBL) was broken: 88%
IB Low 50% extension (IBLX1) was broken: 67%
IB Low 100% extension (IBLX2) was broken: 44%
IB Low 200% extension (IBLX3) was broken: 20%
This happens on about 41% of all days analysed.


The data can be analysed in many other ways. Interesting things to look at might be the proportional of days where both IBH and IBL are broken, or when both IBHX1 and IBLX1 are broken.

IBH and IBL both broken (by 10 pips or more): 21%
IBHX1 and IBLX1 both broken (by 10 pips or more): 5%
IBHX2 and IBLX2 both broken (by 10 pips or more): 1%

Analysis of FTSE100 Initial Balance

Doing some analysis of the FTSE100 Initial Balance. Data is from September 2011 until Jan 2016, and it based on London session data (8:00am until 16:30pm).

The Initial Balance (IB) is the 1st 60mins of trading, and this IB can be a good indicator of the full range for the day. Furthermore, IB extension (at 50%, 100% and 200% above/below the IB) are also postulated to be useful targets or reversal points.

So, lets look at this...

Total number of days analysed: 1103
Average High-Low range: 72.9, standard deviation 35.6
Average IB range: 36.5, standard deviation 18.6
IB High (IBH) was broken: 54%
IB High 50% extension (IBHX1) was broken: 29%
IB High 100% extension (IBHX2) was broken: 17%
IB High 200% extension (IBHX3) was broken: 4%
IB Low (IBL) was broken: 54%
IB Low 50% extension (IBLX1) was broken: 34%
IB Low 100% extension (IBLX2) was broken: 18%
IB Low 200% extension (IBLX3) was broken: 6%

Note 1 : Data obtained from modified SessionLines indicator
Note 2 : Analysed using Excel
Note 3 : Test for a "break" allowed for 5 pips leeway, e.g. had to break the IBH by 5 pips to be counted

What can we say from this data?
  1. Well, the values for high side and low side look very similar, this is good, it means the market was pretty symmetrical over this period.
  2. The 50% IB extension (IBHX1 and IBLX1) was broken very often ~30% of the time, but the 100% IB extension (IBHX2 and IBLX2) was only broken about half as often ~17% of the time, and the 200% IB extension (IBHX3 and IBLX3) was only broken ~5% of the time.

Looking in more detail at the ratio of High-Low range to IB range, I generated the following data.


The % column is the proportion of days that fall into a particular "bin", e.g. for the bin labelled 1.25 it counts those days where the R/IB ratio is between 1.00 and 1.25. The data is left-skew, with a peak at around 1.5 (i.e. 150%), but also quite high proportion up to an R/IB ratio of 3. Days with R/IB ratio of 1 (i.e. the IB range was approx 100% of the total range) were very few.

This type of data gives a feel for what might be expected in terms of normal ranges, based off the 1st hour of trading.

Further analysis will be performed on the data. One thing I want to assess is the proportion of days that show range extension only on one side of the IB.

On days, when the IBL was not broken (by more than 5 pips), the following stats are obtained...

IB High (IBH) was broken: 83%
IB High 50% extension (IBHX1) was broken: 50%
IB High 100% extension (IBHX2) was broken: 30%
IB High 200% extension (IBHX3) was broken: 7%
This happens on about 45% of all days analysed.

Similarly, on days when the IBH was not broken (by more than 5 pips), the following stats are obtained...

IB Low (IBL) was broken: 84%
IB Low 50% extension (IBLX1) was broken: 57%
IB Low 100% extension (IBLX2) was broken: 34%
IB Low 200% extension (IBLX3) was broken: 5%
This happens on about 44% of all days analysed.

Now, this is starting to get interesting. If IBH (or IBL) is not broken (by more than a few pips) then the chance of breaking the other side goes up very significantly.

The data can be analysed in many other ways. Interesting things to look at might be the proportional of days where both IBH and IBL are broken, or when both IBHX1 and IBLX1 are broken.

IBH and IBL both broken (by 5 pips or more): 16%
IBHX1 and IBLX1 both broken (by 5 pips or more): 3%
IBHX2 and IBLX2 both broken (by 5 pips or more): 1%

This is quite interesting, only 3% of days have a break both of the 50% extensions, and only 1% break both 100% extensions.

It's all a numbers game when it comes down to it.

I'd like to do a similar analysis on the DAX, but this is quite time-consuming work, maybe later.

Thursday, January 28, 2016

Session Lines Release 9

I've just uploaded Release 9 for the Session Lines indicator. It's only a small change...

Release 9:
  • Ability to show “Regression Channel” or “Regression Bands” separately

The regression (band or channel) is started at the session open, and continues until the session close. The centre line is the linear regression over this period, and the channel or bands are drawn at +/- 2 standard deviations away from the centre line.

More information on the Session Lines indicator can be found here.

Wednesday, January 27, 2016

Oil rally, stocks rally but then fall-back after FOMC

From the Multi-Compare indicator...


For more information on the Multi-Compare see here.

S&P500 rally and fade

S&P500 rallied earlier in the day following the oil rally (caused mainly by rumours, despite higher crude inventory number - go figure). FOMC later knocks it back.

Some nice levels indicated by the SessionLines indicator.


More information on the SessionLines indicator can be found here.

And here's a trade I just made as it was breaking lower.

Currency Strength (27/1/2016)

As expected on FOMC day, forex is fairly range-bound.

Tuesday, January 5, 2016

RANsquawk FOMC minutes Preview: Looking for the clarification of "gradual"

Update on Currency Strength, Multi-Indicator, Multi-Instrument Compare

I previously reported issues with these indicators (Currency Strength Meter, Currency Strength Bars, Multi-Indicator and Multi-Instrument Compare) which caused by issue with FXCM historic data server. It seems that this issue is now fixed, and it appears to be working again.

You might have to clear your "quote storage". Here are the instructions that I received from FXCM support guys...

1. Delete the History folder in C:\Program Files (x86)\Candleworks\FXTS2\QuoteStorage
2. Reopen TS2

A further advice given was to restart internet router and reboot the PC.

Note: When you re-open TS2 and MarketScope and your charts, it will take quite a long time to load the first time (because it is reloading all the instruments that you previously had cached locally). In fact, I even had some charts fail to load the 1st time, but if you wait for a while, and the re-load your chart layout (or re-start MarketScope) eventually it should load ok.

Monday, January 4, 2016

Currency Strength Meter, Currency Strength Bars and Multi-Indicator

Just a "heads-up" on these indicators. Since the start of the year, there seems to be some issue with the data from FXCM server, and these indicators are failing to load historic data from before the start of year (anything before 1/3/2015 20:30 UK time seems to cause problems). This is stopping the indicators from loading properly.

I have raised a bug report with FXCM and FXCodeBase. In the meantime I am looking for some kind of workaround - but it appears to be issue at their side, and there doesn't seem much that I can do about it. Sorry for the inconvenience.