Posts

The V2 Trading System – An Algorithmic Trading System

Philipp,

Thank you for your email and for your detailed trading system questions.  I sincerely appreciate your professional approach to our trading solutions and I trust you understand that I will not be able to disclose any proprietary Algorithmic Trading System design logic or functions to you at this time.  I will be happy to answer your questions to the best of my ability though.

Please find detailed answers to your trading system questions, in order, below :

Trading System Customer Concerns:

Trading System

Trading System

Algorithmic Trading System QUESTION – please sent detailed trade report from Trade Station (as I think you use) showing all the stats from walk forward analysis (not in-sample, only out of sample)I have included the current AlgoTrades V2.  These statistics are based on a $100k USD initial capital and are LIVE, FORWARD TESTED with the systems firing trading actions into this account.  This is the best form of LIVE testing we can accomplish at this time.  The reason the data starts on July 13 is because we reset this testing account data after making a minor change to the quantitative analysis model within the V2 application.  Over the past 4+ months, we have been running V2 and resetting this account continually as we continued to debug the automated trading application.  As you stated you are developing in AMIBroker and I would assume TradeStation, I’m sure you understand the process of developing these types of trading systems.  Build it, fire it up and debug it till it’s perfect.At this point in time, we will not be resetting this account until September 1st, when we launch the new V2 application for all our users.  We are using this account to debug and test the AlgoTrades V2 Trading system and to show our subscribers how well the system is performing.  Please allow me to go over some of the statistics with you.

With a $50k starting account (over a 10-day trading period : 2 weeks) :
ROI on the account would be : 13.4% : just under $7000 USD
Max Drawdown would be : 2.2% : About $1200~1500 USD
Accuracy ratio (winners vs. losers) : is currently 41.7%
Profit Factor : is 10.0 : 1 with a total of 12 trades.

Now, a breakdown of these values would equate to the following :
Average Winner : $700
Average Loser : $125


The Accuracy ratio will improve over time (as we have seen in our prior testing).  When we reset the account, the statistical data begins to accumulate based on current trading events.  As we build more data, the statistical data will normalize to values that we believe will be consistent going forward.  I would estimate an accuracy ratio of 63.5% or higher going forward.
Given these values, the Sharpe Ratio would be (an estimate) near 4.5 or higher.  The reason for this is that the winning ratio is nearly 6x vs. the losing ratio.We will be sending out these charts on a weekly basis throughout all of August to show our members what they can expect to see from the V2 trading system.  Hope this helps.

ALGORITHMIC SYSTEM TRADING QUESTION – what is the walk forward time frame you have up to now? already a few month?? none at all?We have gone over many years of data with regards to the new V2 system.  This is historical visual analysis based on the trading rules we have established for the V2 trading system.  This is the evidence that we used to substantiate our improvements and development of the V2 system.Since the creation of the V2 trading system, we have been in continual debugging mode for over 5 months.  Because of this, we have reset the testing account many times and have dumped data that was not accurate because of trading system bugs.  Please understand that when we develop these systems, we need to see at least 2~3 weeks of error free trading activity before we begin to have any confidence in the system’s ability to operate for our users.  One minor bug and we start the 2~3 week testing phase all over again.

At this time, we last reset the account on July 13, 2015 because of a minor issue with the “leg limiter” feature.  This issue presented a problem with certain reversal type orders and once we resolved this issue, we decided to reset the account to show our users the REAL OPERATIONS of the V2 trading system.

ALGORITHMIC TRADING SYSTEM QUESTION – what it the ratio of (average $ of winning trades) / (average $ of loosing trade) as I know in actual system this was < 1 (0.6 or similar)Please see my answer to your first question (above).  Should you need further data, I will do my best to provide it for you.
ALGORITHMIC TRADING SYSTEM QUESTION – so you basically have two systems in one: a trend following and a mean reverting system? How do you switch from one to the other?Yes, this is correct.  The primary difference between V1 and V2 is that V2 includes a “Price Rotation Trading System” (not a reversion to a mean system).  Let me try to be clear about this.  A reversion to a mean system is designed to look for price anomalies (oddities) that may happen and then attempt to trade the price swing back to the “mean” (or average price level).  These type of systems have inherent flaws in their designs because the markets are not always rational and price oddities can happen on a very regular basis.A Price Rotation System is different from a Reversion system in one factor, mostly…  that price does not have to be within an anomaly (or oddity) level to be traded.  Think of it like this, price is always trending – even in sideways markets.  Oddities are when price has trended into unusual or unexpected ranges.  Price rotation is when price is within usual or expected ranges, but is simply rotating and not trending directly.

From experience, if you are trying to develop a reversion to a mean system, be very cautious of extended trending and increased volatility.  These factors, and others, will destroy the success you may be seeing within certain testing periods for historical backtesting.  Additionally, if you quantify your analysis modeling engine well, extended trending and volatility may produce incredible results if you can develop and identify a price rotation model to deploy.  With a price rotation model, volatility and trending move in an out of trending vs. rotation seamlessly.

How do we switch from one to another?  That is something we can’t disclose to you as it is part of our quantitative analysis modeling solution and is core to the success of our system.  The only thing I can disclose to you is that our modeling systems migrate from trending to rotation models seamlessly.  Meaning, we can be in long trending mode and move into price rotation mode without giving up an inch with our systems.  Sometimes this migration works in our favor, sometimes we take a small loss – but that is trading.

It sounds like you are trying to develop your own system and I can tell you that I have a great team that works with me and well over 35+ years of trading system design experience within my team.  Without my team of programmers, none of this would be possible.  My trading experience and knowledge translates into working automated trading systems easily.  It is not easy to create a system like this and I understand you might be a bit frustrated with the V1 system and the loss.  Please take a cautious look at the results of the V2 system and pay attention to the future emails from us for more details.

We honestly value your participation in our trading solutions and value YOU as a client.  That is why we are giving you 6 months FREE extension with the AlgoTrades V2 system.  We want to earn your business and continued trust.

Respectfully,