I have been trading for approximately 3 years now, I typically trade for 3 months then take a 3 month break to deepen my sophistication. Initially I started out flapping about making string after string of discretionary trades not making any real money. I progressed from there not into technical charting analysis but order book analysis and masses of reading around market mechanics. Until recently I have focused on trading illiquid Oil and Gas stocks listed on the FTSE AIM market in the UK. These securities have numerous volatility expansions as micro-caps attempt to make large oil discoveries. Which makes them interesting to trade. And I have gradually become proficient in this arena, by learning how to read information value from order book analysis. I have also put a significant amount of work into building a system which tests the liquidity within the yellow strip as the London Stock Exchange allows hidden limit orders to be used on the book. The liquidity imbalance at the yellow strip along with the arrival rate of market orders at the bid and ask is what determines when the price will tick up and down. However as I am a retail trader I can only do this on fairly illiquid securities with fewer than 2,000 trades/day, otherwise I lack the processing power to keep up with the market. I am looking at expanding my system to include technical indicators and stock scanners to broaden my horizon and expand into new markets. I am currently looking at using MultiCharts and ProQuote as a starting point for developing this additional layer to my system. I am also looking at moving towards more automated execution. My question here is regarding backtesting, I am looking for a platform that is robust and effective for backtesting and forward testing. I have 12 months allocated to building and developing this new system which I intend to use technical indicators to scan the markets for the underlying microstructure dynamics with which I am familiar and know how to arbitrage. Does anyone here have experience with this kind of thing? (I'm basically trying to seed a debate on the best way for me to proceed) No time wasters please
Wow, my post finally made it through moderation. Although I missed out the main part which was the Risk Modeling as referred to in the title. Basically I want to employ a number of sub-strategies simultaneously, and each of these sub strategies would be more suitable to either ranging, trending or breaking out markets. Since trending strategies do so poorly in ranging markets and ranging strategies do so poorly in trending markets, I am looking to employ some performance based metrics to allocate trade weights to each strategy, so in a trending market my ranging strategy would be losing more trades than in a ranging market and this would be recognized in the performance metrics and my ranging strategy would be allocated smaller trade weights until it became profitable again (the market started ranging again). Likewise in a ranging market the trending strategy underperforms so when the trending strategy began to underperform the system would assume the market was ranging and allocate smaller trade weights to the trending strategy. My question is; Rather than use average (pts) win and average (pts) lost and number won and number lost. I am looking for something a little more nuanced, perhaps taking a distribution curve of both win side and loss side of the systems performance with a rainfall count and see how the shape changes from trending to ranging and ranging to trending for both the ranging and trending strategies. This way I can use a scalpel rather than a hack saw, and cut the trades which are at greatest risk of underperforming in an unfavourable market. Does anybody here have any experience with something similar? Thanks,
stu, Elitetrader is extremely busy this AM concerning general traffic and moderation.. Pardon the interruption. ES Wow, my post finally made it through moderation.