Systematic Trading. Robert Carver MTA webcast, 2nd March 2016

Transcription

1 Systematic Trading Robert Carver MTA webcast, 2nd March 2016

2 Why trade systematically? Mistake #1: Overfitting Mistake #2: Too much risk Mistake #3: Trading too often

3 Why trade systematically? Mistake #1: Overfitting Mistake #2: Too much risk Mistake #3: Trading too often

4 Why trade systematically? Mistake #1: Overfitting Mistake #2: Too much risk Mistake #3: Trading too often

5 Why trade systematically... Human biases Transferable Repeatable Automation is possible (lower running costs) Large portfolios are easier more diversification Back testing: Confidence it would have worked in the past (may in the future) Understanding of likely behaviour Knowledge of performance / behaviour envelope

6 Bias: Which brain do you trade with? System one (fast, instinctive, emotional) System two (slow, effortful, infrequent, logical, calculating)

7 Bias: Why trend following works...

8 Why trade systematically... Human biases Transferable Repeatable Automation is possible (lower running costs) Large portfolios are easier more diversification Back testing: Confidence it would have worked in the past (may in the future) Understanding of likely behaviour Knowledge of performance / behaviour envelope

9 Why trade systematically? Mistake #1: Overfitting Mistake #2: Too much risk Mistake #3: Trading too often

10 Mistake #1: Over fitting

11 Types of overfitting Explicit: Fitted in the backtest Implicit: Discarded, not in the backtest Tacit: Included in the backtest with tacit knowledge

12 Explicit: Fit out of sample with expanding window

13 Explicit: Use robust fitting techniques

14 Explicit: Use robust fitting techniques

15 Implicit and Tacit Reduce the number of alternatives considered Every model / variation dropped will inflate past performance Reduce iterations: Don't try and polish coal Don't improve (overfit) if performance okay Be aware of tacit overfitting no solution (?)

16 Why trade systematically? Mistake #1: Overfitting Mistake #2: Too much risk Mistake #3: Trading too often

17 Mistake #2: Too much risk What percentage of capital is safe to put at risk on each trade?

18 Mistake #2: Too much risk What percentage of capital is safe to put at risk on each trade? Answer: it depends.

19 Determining bet % Optimal target portfolio risk: (Realistic) expectations of Sharpe Ratio Type of trading / holding period Skill Diversification Backtest degradation Higher moments (skew, kurtosis ) Attainable, given leverage limits Safe, given leverage required Investor comfort zone Portfolio size (average # positions) Holding period

20 Calculate correct bet size Holding period SR: days 9 days 17 days 1.5 months 3 months 26 months Average number of positions % 6.3% 9.4% 12.5% 25.0% 31.3% 1.6% 3.1% 4.7% 6.3% 12.5% 15.6% 0.6% 1.3% 1.9% 2.5% 5.0% 6.3% 0.3% 0.6% 0.9% 1.3% 2.5% 3.1% 0.2% 0.3% 0.5% 0.6% 1.3% 1.6% Holding period SR: days 9 days 17 days 1.5 months 3 months 26 months Average number of positions % 12.5% 18.8% 25.0% 50.0% 62.5% 3.1% 6.3% 9.4% 12.5% 25.0% 31.3% 1.3% 2.5% 3.8% 5.0% 10.0% 12.5% 0.6% 1.3% 1.9% 2.5% 5.0% 6.3% 0.3% 0.6% 0.9% 1.3% 2.5% 3.1%

21 Why trade systematically? Mistake #1: Overfitting Mistake #2: Too much risk Mistake #3: Trading too often

22 Law of Active Management Pre cost Typical future Spread bet month 1 week 1 day Half a day One hour

23 Law of no free lunch Pre cost Typical future Spread bet month 1 week 1 day Half a day One hour

24 Some lunches very expensive month week 1 day Half a day One hour Pre cost Typical future Spread bet

25 How do we get rich? Increase pre-cost performance Same stuff, done quicker Different stuff, done quicker Better stuff, at the same speed Reduce trading costs Reduce commissions Increase AUM Trade more Provide liquidity Reduce slippage Patience Provide liquidity

26 Do same stuff quicker Theoretical Actual month 1 week 1 day Half a day One hour

27 Do different stuff IMC / Virtu Buffet Years Months Minutes Seconds Milliseconds

28 Do better stuff at same speed Signal diversification Market diversification (We've already dealt with Time diversification...)

29 Better stuff - Signal diversification

30 Better stuff - Instrument diversification

31 How do we get rich? Increase pre-cost performance Same stuff, done quicker Different stuff, done quicker Better stuff, at the same speed Reduce trading costs Reduce commissions Increase AUM Trade more Provide liquidity (rebates) Reduce slippage Patience Provide liquidity (capture spread)

32 Capturing the spread Fast Slow (with algo) Trend follower Pays Both (net flat) Mean reversion Captures (normally) Pays (if stops hit) Both (net flat)

33 My book is available from: systematictrading.org

34 And / or, you can read my blog: qoppac.blogspot.com

35 Spreadsheets: systematictrading.org/resources

36 My open source python backtesting software: qoppac.blogspot.co.uk/p/pysystemtrade.html OR github.com/robcarver17/pysystemtrade

37 My book, and other resources, are available from: systematictrading.org Blog, and python code links: