Quantitative Trading System For The E-mini S&P

Transcription

1 AURORA PRO Aurora Pro Automated Trading System Aurora Pro v1.11 For TradeStation 9.1 August 2015 Quantitative Trading System For The E-mini S&P By Capital Evolution LLC Aurora Pro is a quantitative trading strategy designed to trade the E-mini S&P futures market. The system employs daily mean-reversion method and breakout method to produce enhanced absolute and risk adjusted returns. The system uses the following key design principles: 1. Market regime switching 2. Volatility adaptive parameters 3. Both symmetrical and non-symmetrical trading logic Aurora Pro has performed well on both out-of-sample historical data and on live market data. The system is designed for the retail trader who wishes to control risk while maximizing returns. This is accomplished by utilizing Aurora Pro s built-in position sizing model. The software is written in EasyLanguage and will execute on the TradeStation platform. This document will provide you with both historical performance results as well as results generated on live market data. 1

2 Two Strategies In One Trading System Mean Reversion and Breakout Aurora Pro is really two trading systems combined into one. The system utilizes both a Mean Reverting strategy and a Breakout strategy. Each strategy can be traded individually or in parallel. Utilizing both strategies allows Aurora Pro to enter shortterm day trades or hold trades over several days. Having two unique trading strategies helps produce a robust, diverse and effective trading system. All trades are executed on a 5-minute chart while most of the trading signals are generated from a daily chart. Thus, Aurora Pro uses multiple timeframes to both open new trades and to manage those trades. Together these two trading strategies and their trading principles lead to Aurora Pro demonstrating robust performance over the past 12 years. Utilizing two trading strategies helps smooth the equity curve as it brings diversification to your trading. Common Attributes Between Both Strategies Include: Uses two timeframes (5-minute chart, daily chart) All trades executed on a 5-minute chart Long only Trades the E-mini S&P 2

3 The Breakout Strategy A Day-Trading Strategy This strategy utilizes volatility adaptive profit and stop loss targets, and symmetrical trading logic to profit in both bull and bear markets. There are no user input parameters to modify. This strategy makes no distinction between bull or bear markets. The breakout level at which a new position is opened is based upon the prior day s price action on the daily bar chart. A stop loss is also determined based upon the same prior day's price action. This results in both a dynamic profit target and stop loss that adapts to the ebb and flow of market volatility. Breakout trading systems on the E- mini S&P are often more difficult to trade since the market tends to exhibit a mean reverting bias. However, utilizing both a simple trend filter and momentum filter applied to the daily chart, Aurora Pro is better able to capture those days when a strong bull trend will take place during the regular trading hours. The breakout strategy is ideal for clients with small trading accounts such as $10,000. The breakout strategy utilizes tight stops and lets profits run as it attempts to capture break-away markets. 3

4 The Mean Reversion Strategy A Swing Strategy Utilizes market-regime-switching, volatility adaptive profit targets and nonsymmetrical trading logic to profit in both bull and bear markets. Several input values are available for the user to modify. On the macro level Aurora Pro will first determine the overall market regime: bull or bear. Marketregime-switching allows Aurora Pro to adapt its trading to the market regime. The behavior of market participants during bear markets can be significantly different than the behavior seen during a bull market. Thus, Aurora Pro will modify its trading parameters to account for the change. This ability results in nonsymmetrical trading parameters to take advantage of the psychology difference of market participants during the two market regimes. Finally, market volatility is taken into account when determining the profit target which allows Aurora Pro to adapt to a changing market. The E-mini S&P futures market has demonstrated mean reversion characteristics for the past decade. Our own studies published have supported this behavior in the stock index market. Most notably we have demonstrated that buying into short-term weakness during an overall bull market outperforms buying into short-term strength. Strong bullish moves in the stock index market are often followed by sell-offs, not continued strength. On the other hand, sell-offs are often followed by sudden surges in buying. Aurora Pro attempts to capitalize on this behavior by monitoring the daily price activity for sharp pullbacks which are then purchased in an attempt to profit on the sudden snap back to the up-side. We define a bull market when price is above its 200-day Simple Moving Average (SMA). 4

5 POSITION SIZING Aurora Pro Automated Trading System Position Sizing & Risk Determine Your Profit Potential vs. Risk Tolerance Aurora Pro is capable of two different methods of position sizing. First, is a fixed number of contracts. Second, is a Percent Risk Model 5

6 Fixed Number of Contracts Per Trade The default configuration for Aurora Pro is to trade a single contract per trade setup. This simply means for each trade the same number of contracts will be traded. While the default setting is one contract, you may trade as many contracts as you wish. Percent Risk Model The Percent Risk Model is based upon two important factors. First, the maximum dollar amount you are willing to risk per trade as a percentage of your account equity. Next, is the starting size of your trading account. For example, a 2% risk model indicates you are willing to risk no more than 2% of your account on any given trade. This is a common position sizing model. Keep in mind a position sizing model is critical for long-term equity growth and maximizing returns. It is advised you study position sizing as an important factor in your trading. Such a topic is beyond the scope of this manual so if you are not familiar with position sizing, please perform an Internet search on the topic. 6

7 How Big Does My Account Need To Be? This is a good question. To answer this question we need to know... What is your account size is. What you re willing to risk per trade. What is the stop loss value for Aurora Pro. Aurora Pro currently uses a variety of stops and some of those stops are dynamic. Analyzing the historical performance of Aurora Pro for the past 12 years we can estimate the average stop loss value to be around $300. However, we know the stop loss value for the Mean Reversion strategy is hard coded at $500. So we should use the larger value of $500. Given this information we can easily compute the account size required to trade this system. Say you wish to risk no more than 2% of your account on any given trade: Let X represent the account size required. 2%(X) = $500 X = $500/.02 X = $25,000 In this example, you should have at least $25,000 in your trading account to trade the Mean Reversion strategy. If you have a smaller account, you could simply trade the Breakout strategy only. The Breakout strategy risks far less dollars per trade. Based upon the historical performance the average losing trade for the Breakout strategy is about $200. Using the same formula above we can estimate what our account balance should be if we decide to only trade the Breakout strategy. 2%(X) = $200 X = $200/.02 X = $10,000 In this example, you should have at least a $10,000 trading account if you are only going to trade the Breakout strategy. One of the leading causes for people to fail at system trading is not being properly funded. Be sure you have a properly funded account based upon a realistic risk level. 7

8 PERFORMANCE Aurora Pro Automated Trading System Aurora Pro Live & Historical Returns In-Sample, Out-of-Sample & Live Returns Unless otherwise stated, all accounts assume $25,000 starting equity. $30 round trip fee is deducted for both slippage and commissions. 8

9 Aurora Pro Backtest - Single Contract Starting Equity: $25,000 Position Sizing: 1 contract per trade Commissions & Slippage: $30 deducted per trade Test Period: January 1, 2000 through June 30, 2016 $70,000 $61,250 $52,500 $43,750 Aurora Pro Backtest - Single Contract $35,000 $26,250 $17,500 $8,750 $0 -$8, Aurora Performance Net P&L $63,275 Profit Factor 2.20 Number of Trades 445 Winning Trades 235 Losing Trades 210 %Win 53% Average Risk Per Trade $ Average P&L per Trade $ Average Winning Trade $ Average Losing Trade -$ Expectancy 0.47 CAGR 6.10% RETURN ON CAPITAL 253% BUY AND HOLD RETURN 56% Expectancy Score 31.69

10 Drawdown Maximum Drawdown (Intra-day Peak To Valley): $3,648 Maximum Drawdown (Trade Close To Trade Close): $2,355

11 Annual Returns

12 P&L Confidence Interval The chart below is a confidence interval analysis of the P&L. The results on this page were based on trading a single contract and assuming a 95% confidence. The confidence interval for winning trades is about $60 while the confidence interval for losing trades is about $20. These two values are used to produce an upper (green) and lower (red) bands above the average trade of $142. Thus both bands remain well within the profitable region (above the zero line). $210 Average Upper Lower $158 $105 $53 $0 Confidence Interval Winning Trades Number of Trades: 235 Average P&L All Trades: $142 StdDev Average Trade: $467 Confidence: 0.95 t-score: 1.97 Confidence Interval: $59.99 Upper Band: $202 Confidence Interval Losing Trades Number of Trades: 210 Average P&L All Trades: $142 StdDev Average Trade: $149 Confidence: 0.95 t-score: 1.97 Confidence Interval: $20.31 Lower Band: $122

13 Monte Carlo Simulations And Equity Confidence In an attempt to determine how likely Aurora Pro will perform into the future based upon random variation of all historical trades, a Monte Carlo simulation can be used. A Monte Carlo simulation (or Monte Carlo Method) is a computational method that uses repeated random samples take from our historical results and uses these to generate possible outcomes. In this case, we provide the historical trades of Aurora Pro as input and the Monte Carlo simulation generates thousands of possible outcomes and computes the likelihood of such an outcome. After thousands of iterations the rate of return is evaluated at specific probabilities. Below is the result of our simulations. The equity curve in blue is a typical looking equity curve generated by Aurora Pro trading a single contract. The Monte Carlo process then generated two bands colored in red (5% confidence) and green (95% confidence). These represents the best case and worse case scenarios of the equity curve.

14 Increasing Return With Position Sizing How much you risk per trade determines both your total return and your maximum drawdown. The following table is a summary of 1,000 Monte Carlo iterations over the entire historical performance of Aurora Pro. The trading system s default settings were fed into the Monte Carlo simulator which included a starting account size of $25,000 and $30 deducted for both commissions and slippage per trade per contract. I also limited the maximum number of contracts to trade at 30. I then tested different percent risk values for position sizing. The percent risk model for position sizing determines the number of contracts to trade based on a percentage of your total trading capital. Each Monte Carlo run has a confidence level of 95%. In other words, for each run, 95% of the 1,000 iterations fell within the given performance parameters. The Return column is the total return on the initial investment. Max Drawdown column is the maximum peak-to-valley closed trade drawdown expressed as a percentage decline from the prior highest peak in equity. The Net Profit column is the final accumulated profit. %Risk Return % Max Drawdown % Net Profit $ , , , , , ,500 First, you will notice as you increase the percent risk your return and maximum drawdown also increase. What you want to aim for is a maximum drawdown that is comfortable for you. Most people find a 20% drawdown or higher difficult to live with. For example, in the chart to the left a 1.25% risk produces a drawdown of 22.1%. This means 95% of the 1,000 Monte Carlo iterations produced a maximum drawdown of 22.1% or less. This implies that 5% of the iterations produced a higher drawdown. A drawdown of 22.1% may be too much for you to emotionally handle. Decreasing the percent risk to 1.0% pushes maximum drawdown to 17.6%. You have to ask yourself, what drawdown can I emotionally handle? In summary, an ideal percent risk value, based upon these results, will most likely fall between % depending upon your personal risk tolerance. Of course there is nothing stopping you with starting out at 1.00% risk and increasing that value over the years as your account and confidence grows. Knowing this information will help determine if you want to trade such a system. Likewise, if you are trading this system and find yourself in a drawdown the above information can help you with your expectations of drawdown and consecutive losers. Drawdowns and consecutive losing trades are part of the trading game. Knowing what is considered a typical drawdown can help you mentally prepare. On the following pages are equity curves and more detailed results of the 1.00% risk and 1.50% Monte Carlo runs.

15 Aurora Pro Trading 1% Risk Starting Equity: $25,000 Position Sizing: 1% Risk per trade Commissions & Slippage: $30 deducted per trade Test Period: January 1, 2000 through June 30, 2016 CAGR 8.43% RETURN ON CAPITAL 165% BUY AND HOLD RETURN 55%

16 Aurora Pro Trading 1.5% Risk Starting Equity: $25,000 Position Sizing: 1.5% Risk per trade Commissions & Slippage: $30 deducted per trade Test Period: January 1, 2000 through June 30, 2016 CAGR 13.54% RETURN ON CAPITAL 563% BUY AND HOLD RETURN 55%

17 LIVE PERFORMANCE Aurora Pro Automated Trading System Aurora Pro Live Returns Trades Taken On The Live Market Back tests are great, but they are not real trading. As you know, executing real trades on the live market is what counts! In this section trades are based upon trades recorded on the live market. I will perform several tests on this data which will provide the most realistic results. As you know, historical backtests tend to be overly optimistic and this is true for Aurora Pro as well. Thus, in this section I will demonstrate the actual results of Aurora Pro and use Monte Carlo simulation to demonstrate and idea percent risk value. Consider the performance values within this sections and other values, such as the percent risk value, as more realistic than the historical backtested results found in the previous section. Commissions and slippage have been taken into account. 17

18 Aurora Pro Live Results - Income The following results are live trades, not back tested results. This equity graph and results assumes net profit are not reinvested. Instead, trade size is always based upon the initial starting capital of $25,000 Starting Equity: $25,000 Position Sizing: 2% Risk Test Period: July through June 30, 2016 $42,000 $39,875 Aurora Pro Live Trades $37,750 $35,625 $33,500 $31,375 $29,250 $27,125 $25, Performance Expectancy Expectancy Score 10 Win Rate 0.44 Net P&L $15,550 Return% 62.20% Trades 138 Average Profit/Trade $ Median Profit/Trade $ Wins 61 Losses 77 %Wins 44% Average Win $ Average Loss -$ Winnings $ $46, Losses$ -$31, Equity Trend 1 Results Net P&L $15,550 Profit Factor 1.50 Total Return 62.2% CAGR 13.1% Drawdown 7.8%

19 Monte Carlo Simulations With Optimized Percent Risk Model In this test I wanted to determine what is the optimal percent risk value based upon return. I also did not want to the maximum drawdown to exceed 25%. Once again, the results were based upon 1,000 Monte Carlo iterations over the actual trades taken by Aurora Pro since trades were tracked in the member s website in July of The trading system s default settings were fed into the Monte Carlo simulator which included a starting account size of $25,000. The results are below. OPTIMAL PERCENT RISK 3.85% TOTAL RETURN AT OPTIMUM 112% MAX DRAWDOWN AT OPTIMUM 24.7%

20 Aurora Pro Optimal Percent Risk The following results are based upon an optimal percent risk based upon the live trading results. This equity graph and results assumes net profit are reinvested. Starting Equity: $25,000 Position Sizing: 3.85% Risk Test Period: July through June 30, 2016 Win-Loss Ratio Net P&L $28,187 Return% % Trades 139 Average Profit/Trade $ Average Win $2,000 Average Loss $1,047 Largest Loss $1, Max Drawdown $13, Net P&L $28,187 Profit Factor 1.33 Total Return 112.7% CAGR 20.8% Drawdown 24.7%

21 FAQ Will there ever be a conflict when trading two strategies in Aurora Pro? Both trading strategies enter at different times so there is no conflict. Furthermore, if a trade is open no other entry signals are taken. This ensures only one signal will be taken at a time. What platforms is Aurora Pro available for? Currently Aurora Pro is available only on TradeStation. Will Aurora Pro work on other markets? In short, yes. The system was designed for the E-mini S&P. However, it will function on other stock index future markets such as the E-mini DOW or E-mini NASDAQ. You will find the performance on these alternative markets not nearly as effective as the E-mini S&P yet, both remain positive. The E-mini DOW performs better than the E-mini NASDAQ while the breakout strategy seems to perform better than the mean reversion strategy. In the end, because Aurora Pro remains profitable on other markets, this is a testimonial to its robust nature. I attached copies of the performance report for the NQ and YM markets to this . Has Aurora Pro been traded live? Yes, within this report on page 9 you will find the live results of Aurora Pro. Is seasonality factored into Aurora Pro's trades? No. Aurora Pro does not take into account seasonality. Does Aurora Pro use limit orders? No, Aurora Pro uses market orders to enter and exit trades. Can I trade with only $10,000 Technically, yes. This is recommend only for the Breakout Strategy. Thus, you will need to disable the Fade strategy portion of the system. Please review the User Manual on how to do this. How big should my account be? To properly trade both the mean reversion and breakout strategy it is recommended you have a trading account of $25,000 or more. Is the source code available? Not at this time. I hope to make it available in the near future!

22 Aurora Pro Automated Trading System AURORA PRO Aurora Pro is a fully automated quantitative trading strategy for the retail trader. Questions Comments: sales@capitalevolutionllc.com