Backtesting Trading Strategies
Can someone who's familiar provide a somewhat detailed walkthrough of how firms backtest trading strategies? Just curious to know the process by which a hedge fund or trading desk would decide to pursue a particular strategy.
create strategy, load into backtesting software.
software forces through historic market data at accelerated speed, and says how the strategy would have resulted.
DOES NOT ACCOUNT FOR SLIPPAGE (changes in the market caused by the actions of your strategies).
Personally, I see backtesting as a bug finder and an indicator of whether it will fail, not whether it will succeed.
I agree, you can code basic scaplers and backtest them, they will be profitable on the simulation but it is highly unlikely they would be under real-life conditions. At least thats what i read when searching to code expert advisers.
depends solely on the correlation of your simulated tick curve(est. model) to the true model. I run a model that is very realistic, so the data is about as accurate you can get to the real market. Also depends on how your algorithm executes the trades.
trazer985 is mostly correct, though he is misusing the term "slippage" when he means "market impact". Backtesting frameworks can model market impact as well.
Backtesting Macro Strategies (Originally Posted: 12/12/2011)
I interviewed with a stat arb shop once and was asked how I backtested my strategies. I found this to be a tough question to answer because I trade macro with a days to weeks holding period.
When I put on a position, I look at some key things: dollar index and major FX crosses, level and price action of key equity indices, yield curves for USD JPY and EUR, credit spreads, price level and price action of key commodities (oil, copper, etc.). But on top of this I look at newsflow. Trading macro, I feel that I am not trying to play off technical features of price action but playing psychologist to the market. And it has worked well for me; I have beat index by a healthy margin over the past 36 months.
But I just don't see how I can backtest this. When in the last 30 years, which is a good backtesting window, can you point to another massive currency breakup of a massive economic zone, for instance? Or a balance sheet meltdown on the scale afforded by the implosion of the CMBS market?
From 2008 on we have had so many exceptional, almost one-off events what useful information do we get out of backtesting? Serious GIGO problem here, at least for how I trade.
Yeah, it would be very tough to backtest macro strategies. I work on an equity stat arb desk, and all of our trades are done on an intraday basis. So backtesting is quite easy. We run python scripts to do backtests on a 1-3 year timeframe for around 2000 equity pairs, using a combination of technical parameters. I also do shorter-term backtests as well if I believe that a fundamental relationship between a given pair has changed during that period.
Facilis eos vero et dolores qui. Rerum eum corporis et dignissimos doloribus dolor tempora.
Voluptatum suscipit quia impedit nemo beatae. Sint consectetur harum delectus consequatur ut ut. Nobis omnis iste ratione.
Beatae neque velit delectus eum quia odit aut. Sit molestiae sed deleniti commodi ut.
Quia tempore minus architecto. Fugit eum porro enim culpa qui rem. Enim necessitatibus aspernatur quia possimus ipsum molestiae quas. Et ut ea similique fuga recusandae ut quia numquam.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...