Money clearly flows into

Size: px

Start display at page:

Download "Money clearly flows into"

Aubrie Shields
5 years ago
Views:

1 TRADING STRATEGIES Controlling risk in a seasonal strategy The end-of-month trade can be improved by paying closer attention to its downside risk. BY EMILIO TOMASINI AND URBAN JÄKLE Money clearly flows into the stock market during the end of one month and the beginning of the next. Over the past 11 years, for example, the S&P 500 index gained an average of 0.24 percent on the first day of the month, closing higher 63 percent of the time. By contrast, the index s average close-to-close move for all days was negligible less than 0.05 percent and the odds of a higher close were around 52 percent. What s the best way to exploit this bullish pattern? Of course, you could buy on the last day of the month and sell one day later. This approach isn t necessarily a bad idea, but it ignores risk; when the market can swing roughly 10 percent in either direction in a single day, as it did in October 2008, adding stoploss rules is a prudent step. The following analysis begins by assessing the simple first-day-of-themonth buy rule and then examines the best time to exit the market and how to improve results by adding profit targets and stop-loss points. The discussion also focuses on how the system has reacted in different market phases bull and bear periods and volatile and calm stretches over the past two decades. The basic approach has been tracked by Futures Truth, an independent evaluator of trading systems, for more than five years. This test uses S&P 500 futures TABLE 1: KNOW WHEN TO HOLD, FOLD Holding period Average profit Total net profit in days per trade ( ) 1 $ $157,812 2 $ $157,875 3 $ $129,062 4 $ $154,675 5 $ $126,787 6 $ $56, $ $60, $ $76, $ $37, $ $70,112 The largest profits emerged when the trades were held less than five days. FIGURE 1: TRADE EXAMPLES The S&P 500 tendency to jump during the first week of the month resulted in three of four profitable long trades in early May 2009 ACTIVE TRADER

2 Strategy snapshot Strategy: End-of-month system. Market: Any stock index futures or ETFs. Logic: Exploit the market s tendency to rally around the first of the month. Rules: Buy the market on the last day of the month. Sell up to four days later. Money Add stop loss of 1 percent, profit management target of 4 percent,trailing stop of (for S&P futures): 2.2 percent, and one-day volatility stop. FIGURE 2: EQUITY CURVE BASIC SYSTEM Without any exit rules, the basic system suffered large losses in the bear market and nearly fell apart in TABLE 2: PERFORMANCE BASIC SYSTEM Total net profit $124, Gross profit $538, Gross loss -$414, Profit factor 1.3 Total number of trades 233 Percent profitable 60.94% Avg. profit $ Avg. winner $3, Avg. loser -$4, Ratio avg. win / avg. loss 0.83 Largest winning trade $15, Largest losing trade -$29, Largest winner as % of gross profit 2.91% Largest loser as % of gross loss 7.01% Net profit as percentage of largest loss % Max. consec. winning trades 8 Max. consecutive losing trades 5 Avg. bars in total trades 5 Avg. bars in winning trades 5 Avg. bars in losing trades 5 Return on initial capital % Annual rate of return 4.08% Buy-and-hold return 7.20% Account size required $85, Total commissions $34, Percent of time in market 19.49% Max. drawdown (intraday peak to valley) -$99, Net profit as percentage of drawdown % The basic strategy earned $124,225 (124.2 percent) overall with a 61-percent win rate and a profit factor (gross profit/gross loss) of 1.3. But its risk was unacceptable as system equity dropped nearly $100,000 in (SP), but it can also be applied to any stock-index futures contract or exchangetraded fund (ETF). The method was roughly as profitable in European and Asian stock-index futures as it was in U.S. markets. When to exit? Before testing a complete strategy, let s look at the basic approach s profitability in holding periods of one to 10 trading days. Table 1 lists the system s average profit per trade and total net profit when trades are entered at the close of the last trading day of the month and exited at the close one to 10 days over a 20-year period (Jan. 26, 1989 to Jan. 23, 2009). The most profit was generated by exiting on the second day of the month, with a $680 average trade and $157,875 total profit (157.9 percent). Holding periods up to five days generated roughly comparable profits, but performance dropped in the second week. Based on Table 1, it seems reasonable to hold the trade up to four days, which is the period that will provide the backbone of the strategy we will explore. The trade rules are: 1. Buy the market at the close on the last trading day of the month. 2. Sell the market at the close on the fourth trading day of the month. Figure 1 shows several trades from February to May 2007 in which the longonly system bought the S&P 500 futures at the close of the last trading day of the month and held the trade four days (unless it was stopped out by profit targets or stop-losses, as was the case in early March). All but one of the trades was profitable. Basic performance results Table 2 shows the performance statistics continued on p. 24 ACTIVE TRADER May

3 Trading Strategies continued FIGURE 3: STOP-LOSS DISTANCES FIGURE 4: PROFIT TARGETS The green area shows the range of stop distances that improve the trading strategy s net profit. Stops up to 3 percent enhanced performance, while larger stops weakened the system. Smaller profit targets hurt the system because they didn t give profits a chance to grow. However, larger profit targets 3 to 4.5 percent - improved results TABLE 3: PERFORMANCE REVISED SYSTEM Total net profit $223, Gross profit $458, Gross loss -$235, Profit factor 1.95 Total number of trades 235 Percent profitable 45.11% Avg. profit $ Avg. winner $4, Avg. loser -$1, Ratio avg. win / avg. loss 2.37 Largest winning trade $16, Largest losing trade -$8, Largest winner as % of gross profit 3.61% Largest loser as % of gross loss 3.55% Net profit as percentage of largest loss % Max. consec. winning trades 5 Max. consecutive losing trades 6 Avg. bars in total trades 3.51 Avg. bars in winning trades 4.58 Avg. bars in losing trades 2.64 Return on initial capital % Annual rate of return 5.92% Buy-and-hold return 7.20% Account size required $19, Total commissions $35, Percent of time in market 13.11% Max. drawdown (intraday peak to valley) -$20, Net profit as percentage of drawdown % Adding a series of exit rules and profit targets doubled the system s net profit to $223,013 from $124,225. The average trade increased to $949 from $533, and the maximum drawdown became nearly five times smaller. for trading one S&P 500 futures contract and $100,000 in initial capital over the past 20 years. Commissions and slippage of $75 per trade ($150 round trip) were included. Figure 2 shows the basic system s equity curve over the test period. The basic strategy earned $124,225 (124.2 percent) overall with a 61-percent win rate, a profit factor (gross profit/gross loss) of 1.3, and an average trade of $533. However, large losses weren t uncommon. For example, the average losing trade was larger than the average winner (-$4,555 vs. $3,794), and the largest losing trade was nearly twice as big as the largest winner (-$29,075 vs. $15,650). Finally, the system lost roughly $100,000 as the stock market tanked in Clearly, the system has potential, but its drawdowns were exceptionally large. Let s look at ways to potentially enhance performance by adding exit rules that limit losses without sabotaging potential profits. Optimizing profit targets and stop losses Table 1 shows the approach was slightly more profitable when held just one or two days, and shorter holding periods reduce the strategy s market exposure. On the other hand, a longer holding period (up to four days) offers more flexibility when developing profit targets and exits for the system May 2009 ACTIVE TRADER

4 TradeStation code Before adding arbitrary exit rules, let s examine how different-sized stop losses and profit targets influence the system s profitability. Placing stops a certain percentage below (or above) the entry price works better than simply using fixeddollar amounts, because percentagebased stops adjust to market conditions. Fixed-dollar stops mean the system uses a tighter stop when the index s value is higher and a looser stop when the market is lower. For example, a $1,000 stop in the S&P 500 futures represents 4.00 points. When the March S&P 500 contract (SPH09) traded at 1,000 on Oct. 7, 2008, that stop was 0.4 percent of its value. But when the market traded at 820 on Feb. 10, that stop represented 0.5 percent of its value. The basic system s total net profit since 1989 is roughly $150,000 (without deducting commissions). Figure 3 shows the system s profit according to a stop s percentage size from zero to 5 percent. The green area shows the range of stop distances that improve the trading system s net profit. All stops from zero to 3 percent did a good job. The most effective stop losses were between 1 and 2 percent below each trade s entry point. With the March S&P 500 futures trading between 800 and 900 in January, a 1- percent stop-loss was a loss of around 8 or 9 points ($2,000-$2,250). Now that a 1-percent stop-loss has been selected, the next step is to check how different-sized profit targets affected system performance. Exiting after a trade moves a certain amount in the right direction can boost results, especially in stock-index futures. Without any profit target, the system gained about $200,000. Figure 4 compares the system s profit according to targets from 0 to 6 percent. Profits shrunk when the profit target was below 2.4 percent, and especially when it was below 1 percent. Such small targets don t give the market enough room to develop; too many trades were exited with small profits that would have grown larger if given the chance. continued on p. 26 [LegacyColorValue = true]; {Copyright Urban Jäkle, Some facts about the system: 1) Only end of day data is needed. 2) It produces long entries for any stock index futures market, e.g. S&P 500, Nasdaq 100, Russell 2000, EuroStoxx, DAX, Nikkei 225, Hang Seng, KOSPI, SMI, etc. 3) All trades are exited after a hold period of 4 days or if not stopped out before by any applied exits. 4) This code includes a volatility-based stop, but no other stop-losses or profit targets. { The two inputs } {1. How long to stay in the market: DaysInMarket(4), } { 2. Distance for the range based stop: Percentage(0.65), } Inputs: DaysInMarket(4), Percentage(0.65) ; Variables: StopDistance(0), LiquidationPoint(0); StopDistance= Percentage*Average(Range,8); LiquidationPoint=Close-StopDistance; {Setting the range based stop} { Entry: Go long on the Close of the last day of each month } If (DayOfMonth(Date)=29 AND DayOfWeek(Date)=5) OR (DayOfMonth(Date)=30 AND (Month(Date)=4 OR Month(Date)=6 OR Month(Date)=9 OR Month(Date)=11)) OR DayOfMonth(Date)=31 OR (DayOfMonth(Date)=28 AND Month(Date)=2) OR (DayOfMonth(Date)=28 AND (Month(Date)=4 OR Month(Date)=6 OR Month(Date)=9 OR Month(Date)=11) AND DayOfWeek(Date)=5) OR (DayOfMonth(Date)=30 AND DayOfWeek(Date)=5) OR (DayOfMonth(Date)=29 AND Month(Date)=2) OR (DayOfMonth(Date)=27 AND Month(Date)=2 AND DayOfWeek(Date)=5) OR (DayOfMonth(Date)=26 AND Month(Date)=2 AND DayOfWeek(Date)=5) then Buy( BuyDay ) this bar on close; { And finally the exit } Sell( IniStop ) next bar at LiquidationPoint stop; {The average range based stop} If BarsSinceEntry=DaysInMarket then Sell this bar on close; {Exit each trade latest after 4 days at close} ACTIVE TRADER May

5 Trading Strategies continued FIGURE 5: REVISED EQUITY CURVE VS. S&P 500 However, wider targets between 3 and 4.5 percent improved results. System profit seemed to peak at a 4-percent target, which translated to about 30 to 40 points when the March S&P 500 futures traded between 800 and 900 in January. Both exits not only increase the basic system s profit, but they also help reduce its drawdowns and risk. The revised trading system incorporates these exit rules and adds two others a 2.2-percent trailing stop and a volatility-based stop with a distance below the market of 65 percent of the 8-day average range (highlow). For example, when the S&P 500 traded at 832 on Feb. 11, a 2.2-percent trailing stop was roughly 18 points below the market, and the range-based stop was about 12 points below the market. The trailing stop tracks drops that can occur over several days, but the volatility stop will close a position if the market drops, say, 12 points in a single day. However, that stop won t protect you against smaller losses that occur on four consecutive days, which is why including other types of stops is a good idea. The revised rules are: 1. Buy the market at the close on the last trading day of the month. 2. Exit the position at a stop if the trade loses 1 percent, or Source: esignal The revised system was unprofitable from 1989 to 1996, but it rebounded in the mid-1990s as the bull market gained momentum, a trend that reappeared in Also, the strategy held on to gains during the bear markets of and Exit a profitable trade with a trailing stop if the market drops 2.2 percent, or 4. Exit the trade at the next day s open if the market drops 65 percent of its 8-day average true range in a single day, or 5. Exit the trade at the close on the fourth trading day of the month. Table 3 lists the revised system s historical performance over the past two decades. The exit rules nearly doubled the strategy s profit to $223,013 from $124,225, its profit factor jumped to 1.95 from 1.3, and the average winning trade was twice as big as the average loser ($4,321 vs. -$1,822, respectively). Finally, the maximum drawdown fell to $20,650 from roughly $100,000 under the basic system. Results in different market phases Figure 5 compares the revised system s equity curve (top) to the S&P 500 index (SPX, bottom) since It shows how the trading system s profitability has changed with different market phases. For example, the system was unprofitable from 1989 to 1996 because the market s point values and volatility were too low, so the system couldn t offset slippage and commission costs (1). However, when the bull market picked up steam in the mid-1990s, the revised system produced strong gains (2), which were repeated in the subsequent bull market from 2003 to 2007 (4). Interestingly, the trading system maintained or even increased profits during the two major bear markets from 2000 to 2003 (3) and 2008 (5). Remember that the S&P 500 dropped in half during these periods, and the system gains ground only if the market climbs at the beginning of the month. Figure 6 compares the gains and losses of the revised system s individual trades to the S&P 500 index. The range of trade results widened in volatile markets and 26 May 2009 ACTIVE TRADER

6 FIGURE 6: INDIVIDUAL TRADES VS. S&P 500 narrowed when volatility fell. In volatile times, the biggest winning trades reached $10,000 to $15,000. But risk of overnight down gaps is also significant. After buying the market at the close, you hold the trade overnight and sometimes over weekends and holidays. Thus, if the market drops sharply, a 1-percent stop-loss roughly 8 to 9 points or $2,000 in mid-february won t limit your maximum losses. During the back-testing period, the S&P 500 futures occasionally dropped overnight, causing losses of more than $5,000 with a single trade. One alternative is to trade the E-Mini S&P 500 futures that have a point value of only $50 five times smaller than the big S&P 500 contract. Another idea is to buy puts and hold them the entire month while trading the system only at the beginning of each month. Finally, the initial system s equity curve was much smoother and consistent when it simply held the trade for one instead of four days (not shown). Adding unique exit rules to this shorter system might be more promising than the four-day system described here. For information on the author see p. 8. Source: esignal As expected, the range of individual trades widened in volatile markets and narrowed when stocks settled down. In volatile times, the system posted gains of $10,000 to $15,000, but losses sometimes exceeded $5,000 if the S&P 500 plunged overnight. How the ISM report impacts the stock market Active Trader, April The market s response to the monthly ISM manufacturing report is quite interesting because it s released on the first trading day of the month. Discover how the S&P fared surrounding this event over the past seven years. Ben Warwick: Taking the quantitative view Active Trader, February An interview with money manager Ben Warwick. His approach is to trade the market s reaction to certain events such as economic report releases. Options Strategy Lab: The May-October effect Options Trader, July It is well known the overall market has tended to perform better from October to May than from May to October. The adage sell in May and go away has received a lot of attention in the financial press this year, and the recent market downturn seemed to confirm the validity of this strategy. This seasonal system uses options on the S&P 500 index to take advantage of this phenomenon. Related reading Random chance of market reality? Active Trader, December How do you know whether the market patterns you find are the real thing or just smoke and mirrors? Several statistical studies can be used to separate valid market events from random occurrences. Find out what these tools reveal about seasonal tendencies in the stock market. Trading System Lab: May-October system Active Trader, July What if you had to make only two trades the entire year one to cover short and go long and another to sell your longs and go short? This is exactly what the May-October system does. Mark your calendars: Stock market seasonality Active Trader, September This article discusses and tests theories of investing in some months, but not in others. You can purchase and download past articles at ACTIVE TRADER May

Quantitative Trading System For The E-mini S&P

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