FX TRADING SYSTEMS Testing the rules
So where do you start? Is there anything to be learned from the traditional quant approaches in equities? Do average prices have any
significance? What is the
trade drawdown and TCA? What about strategies in other asset classes -
Statistical
A rb i t r a g e, Global Macros etc? How will High / Med / Low Frequencies impact the outcomes? Will AI
be useful?
Which one? Deep Learning, B a yesia n I n fe r e nc e, N e u r a l N et w o r k s, Gau s s i a n P r o c es ses, SVM’s?
out the following tasks – Market
Analysis, Market Monitoring and Trade Monitoring. The majority of robots “out there” merge these functions into one – and it gets very messy when things go awry. Keeping the algo with functional borders make
them so much easier to
the market is behaving normally or if any adjustments are expected. Both of these modules then have to feed into your Trading module in order for it to place trades.
If the algo you’re designing or
robot you’re purchasing does not do the above – ultimately it is a waste of time. A tsunami will hit.
While I am not a big fan of using predictive maths or SVMs for defining trading rules, I find the Bayesian discipline very useful indeed for
testing
You have to back test against 8 years of historical data before attempting to apply the algo on a live account
To ensure that you are not falling into the retro-trap you will have to back-test
the method against
12 months of data. If the results are promising after that you really have to back test against 8 years of historical data before attempting to apply the algo on a live account. That’s what professionals do.
Functionally the robot has to carry 78 FX TRADER MAGAZINE April - June 2016
maintain should it prove necessary (for technical reasons – e.g. change of price feed, change of currency cross etc).
Te Analysis Model should operate overnight and provide a volumetric analysis of the market size and also which direction the market will be moving. The Market Monitor runs constantly and indicates whether
and proving the the
rules once rules are
defined. But while the hoards of algo de v elo p e rs
insist on scatter-gun statistics and the analysis of volumetric moves to identify the trade rules the profession remains pretty much like the medieval apothecaries, addressing symptoms and not causes.
Anthony Stephens
Head of Algo Development Torion Joint Venture Group
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95