FXQU_A1 v4.43

I have included some new velocity and acceleration algorithms in the signal detection and money management code.
This has been deployed to some accounts for final testing... V4.43

Fxqu v4.36

Latest version v4.36 was dropped last night.  It contains an amended volatility calculation that takes spread into account.  Recently with the GBP volatility on brexit it has become necessary to capture these spread fluctuations.

FXQU_A1 v4.01

Version 4.01 was deployed on the 14th November.. Ie last week.  The restriction on take profits was removed and the trades are now allowed to hit their TP and SL targets

FXQU_A1 v3.23

Well v1.89 has been upgraded to v3.23 across all accounts.  It is more advanced in most aspects and now covers both mean reversion and counter mean reversion signals. 

FXQU_A1 v3.14

Been a busy month since my last update.  V3.14 is deployed across most accounts and contains significant advances compared to v1.89...

Quartz has dipped into overall negative returns and is struggling with the current type of market movements.  Version 3.14 has much better value capabilities to handle this type of market.. Will see what the forward testing looks like.

FXQU_A1 v2.59

V2.59 was deployed last night.  It contained a minor change to the risk management coding.

No there changes.  NFP tonight and so it will be interesting in any case

FXQU_A1 v2.57

I have deployed v2.57 across all bots except quartz.  It contains an updated stop loss and take profit adjusting algorithm that greatly reduces the number of changes made to TPs and SLs

FXQU_A1 v2.53

A small change to the volatility detection algorithm was made last night and deployed across two of the bots.  One demo and one live prototype account.

The battle between spreads, noise and true price movement makes for an extremely difficult challenge.. And having a robot make decisions influenced by these factors is very difficult.  However the bots never get tired.. They only get it right or wrong.

Quartz v1.89 remains untouched.

FXQU_A1 v2.51

I have updated the risk management code for the prototype bots.. V2.51 is deployed across several accounts.  There is also two tiny timeframe accounts in testing - these cover synthetic timeframes from 2 to 5 minutes. Spread and commission now play a direct part to the profitability of these settings.

FXQU_A1 v2.47

Latest version v2.47 is deployed across all demo and live prototype accounts.  Quartz is excluded.  New features include a smarter risk management algorithm as well as a new adjustable hard stop loss.

Lots been worked on in the background. More to follow

FXQU_A1 v2.42 sniping

I have added a sniping function to the bots so that for any first level only position there is a rapid exit opportunity should the profit equate to a small positive amount.  If this exit action is not taken then the normal algorithm takes over.  This is to reduce overall market exposure and also to increase the broadness of trades a available.

A couple of versions ago I introduced a neighbour limit whereby timeframes scanned can not be open at the same time.  This allows for more timeframe diversity and alslimits market exposure.
V1.89 (Quartz) was not touched.

FXQU_A1 v2.39

Stops were taken out last week and this was a good outcome considering the carnage on markets.

Quartz is still positive overall but August looking likely to be a negative month.

V2.39 has been deployed to a live prototype account for testing

FXQU_A1 v2.36

V2.36 is now running on 3 accounts.. V1.89 quartz has not been upgraded as this is the primary account with 8 months of history.

The robot has evolved into a highly active machine that combats spreads, low noise trending and news events.  I will continue to refine the bot and upgrade the prototypes as needed.

Mid July

No new versions deployed this month to the prototypes.   I am still investigating better ways to build on my current volatility measurement algorithms- all conventional methods are useless.

I have switched one of the demo accounts to look at timeframes between 30 and 60 minutes... Less than 10 minutes appears to be the borderline at which transaction costs eat any profits made due to the high frequency.

FXQU_A1 v1.89 vs v2.20

June will finish down for FXQU_A1 Quartz which is v1.89 of my fx bot.  The latest version is v2.20 and this has a more advanced and accurate set of algorithms to measure and act on volatility. 

The large single direction of movement during June caused the losses and v1.89 has limited ability to handle this.

V2.20 will continue to be tested on a live account for now.. Unfortunately nothing beats live account testing..

FXQU_A1 ... big moves

Been a while since I posted.   June is turning out to be a negative month with large moves in the EUR based pairs and also JPY.  Strong directional moves with low levels of noise or swings has seen Quartz retrace and its Max DD popped over 10%.

Quartz is using v1.89 of FXQU_A1 whilst the latest prototype is using v2.17

Its worth noting that I have demo accounts with exactly the same setup as a live account and the trades are different.  Goes to show nothing beats using a live account to do final testing.

I am still looking at given the robot more ability to make decisions in the markets as seen in the last few days .. particularly in measuring volatility.

FXQU_A1 .. two new prototypes

There was a large JPY pair movement on Wednesday and it was interesting to see how the bot handled the directional movement.   Whilst Quartz is the stable and untouched version since Feb I do have two other prototypes running as well on live accounts.  One of these prototypes has a unique hybrid standard deviation algorithm built specifically to handle significant volatility changes to the FX market.  It handled the JPY movement much better than the other two bot configurations.

When I get a few months track history with the prototypes I will make them visible the public.

FXQU_A1 using stops

The importance of using stops was once again verified.  EURJPY and GBPJPY pairs were all stopped out this morning for a 1.5% loss total on the account. 

An important part of using stops (and take profits) is to have them continuously adjusting to the market volatility...

FXQU_A1 secondary vps down

So the reason for being careful about platform upgrades and following proper process is again evident.  I performed a normal Windows server update on the secondary vps over the weekend and it has fallen over.  After trying numerous different tasks I couldn't even log in to it nor could the platform vendor.

The server has been reprovisioned however this means reinstalling everything.  I hope to switch the vps to a dedicated server in the near future.

FXQU_A1 v 2.07 update

Lots happening in the background. Currently testing v2.07 which has a new volatility algorithm... It is a hybrid version of standard deviation.. Convention was again thrown out the window.

More later

FXQU_A1 v1. 89 vs v2. 04

Its time to verify the most recent version (v2. 04) with the control version (v1. 89).  This will be done with live accounts and on the G7 pairs only.

There is a lot going on in the background and I will update later

FXQU_A1 smallest timeframes

Well it's looking like the smallest timeframes, 10 mins and less, are the threshold for retail high frequency trading for FXQU_A1.  Possibly going to change the time frames for graphite as this account uses these very small timeframes.

FXQU_A1 standard deviation

It's been flat couple of weeks overall for the robots.  Watching the behaviour of the bots I was keen to capture a better measurement of volatility since this has significant impact on the calculations.  I have now introduced using standard deviation into the code to measure the volatility and deployed the new version across all bots except quartz and graphite.  I have also brought FXQU_A1 majors and A2 majors back on line - these are demos.

FXQU_A1 update.. Lowered risk

What a night.. FXQU_A1 Sapphire and Topaz took a hit due to their high frequency scaling algorithm.  I have amended this to bring it down and in line with the graphite and quartz accounts.

Given the volatility over the last few hours the two accounts FXQU_A1 Quartz and FXQU_A1 Graphite handled the movements really well.  Drawdown is still in hand.

FXQU_A1 update.. Crypto.. MAM

Not much to add.. V1.92 deployed to to two robots.. Cosmetic changes only. The crypto currency offering is progressing... Even though cryptos are getting bad press.. Go against the Herd.

Might also start looking a MAM account

FXQU_A1 update... Quiet

OK so all 4 robots going ok... Not sure why I call it 4 robots because it's the one engine driving them all..

I deployed a slight configuration change to the two most active accounts sapphire and topaz to slow them..

Apart from that, my plans to proceed with a new crypto are progressing. More later

FXQU_A1 cryptocurrency

All 4 robots going well... No new versions deployed.

In other news I will be creating a new passive income cryptocurrency that uses the Fx robots to provide daily returns to token holders.  Looking to disrupt the scam filled arena that is currently the lending crypto world.
More later...

FXQU_A1 vs a massive push

CPI data from us was released last night... The number was high was USD pairs took off... 100 pips in a few minutes. The robots had most of their trades closed out as stops or take profits were hit.   It's 12 hours later the markets have reversed however they are directional across the smaller timeframes.   The robots are down for the day but draw downs and losses are minimal.

FXQU_A1 v1.89

All robots are now switched across to v small time frames and they had their first real test yesterday... Jpy pairs moved over 150 pips and most of the robots hit stop losses and dropped 1% to 2%... Draw down was less than 10% which is great.

V1.89 was deployed as it contained a minor change to the entry trigger.

FXQU_A1 small average durations

FXQU_A1 Sapphire is unwinding its exposure to the usd pounding nicely.

However I am now changing the approach of the robots.   I am changing the settings for all of the accounts to enable much smaller average trade durations - dropping from 2 days to 2 hours.   This will help avoid the broker limits being hit.

Three new accounts are being set up for this reason.. FXQU_A1 Quartz, FXQU_A1 Graphite and FXQU_A1 Topaz.  I will be phasing out the original FXQU_A1 Sapphire account once the open trades complete their natural close outs.

FXQU_A1 reset..

So given the constraints of brokers and changes made I have decided to do a reset.  Two new live accounts have been created - one with ic markets and one with AxiTrader.   The AxiTrader account is aimed at attracting subscribers so very low trade counts and low draw down is the aim.   New names for these coming soon.

FXQU_A1 broker limits hit.... Again

Time to adjust the approach with the accounts.  I have hit the trade limits again.

I will changing things up this weekend... New approach... More later

FXQU_A1 update... Smaller timeframes

V1.80 was deployed last night to now give the option to stop new trades opening if the same direction and timeframe was already opened. Also some cosmetic changes such balance, equity etc is displayed.

Tough going at the moment with single direction of the usd pairs.. I have switched my second live account Ruby to very small time frames however the already opened large time frames will need to eventually close out.

FXQU_A1 A2

OK so I plan to switch the second live account (Ruby) over to the A2 model which looks at very small timeframes... This way I will have two accounts covering small and large timeframes.

No new versions deployed..

FXQU_A1 re-factoring

Latest version is v1.78  ...  I had to refactor the engine to cater for broker limits (max trade numbers).  This meant some prolonged testing periods across all the robots - it also meant that a minor bug crept into the deployed code.  This is now fixed however I am monitoring the internally generated messages created during the robot execution window.

FXQU_A2 update

Since mid December the robots have been pushed... Constantly trying to manoeuvre through positions to get the best outcome in a one directional market.

I have changed the FXQU_A2 account to very small time frames as of this morning.

Latest version is v1. 77

I had to recode some data blocks that reduce the level of broker server messaging following a warning from IC markets.... Apparently 2000 hits per day is a lot... Unfortunately that's not a lot for FXQU_A1!

FXQU_A1 update.. Dynamic throttling change

New minor version released v1. 74... This was to all robots.  Key change was a slight mod to the dynamic throttling algorithm.

The robots are currently attempting to recover from the long usd downward run and are positioning up appropriately.

FXQU_A1 update.. Limits hit

With the strong directional decline of the usd pairs I noticed some trades were not completing. Further investigation showed the robots were hitting the Max allowed number of trades per account (in line with broker guidelines) . I have changed the code accordingly and have deployed the latest version 1.72 across the robots. 

Overall the next week or two the new version will be fully flushed through the trades. 

FXQU_A1 2018 .. 1st update

Happy New Year .. the last week of December 2017 saw a general dumping of USD and this has meant a build up of trades for most USD pairs.   

A minor new version was deployed to all Robots (v1.58) - it includes an updated, stricter, dynamic throttling algorithm.

Also looking at including the first crypto-currency pair : BTC/USD. but will check it over first.

Thats all for now .. 2018 here we go.