| StockFetcher Forums · Filter Exchange · \"HIGH PROBABILITY ETF TRADING\" BY LARRY CONNORS - GET YOUR FILTERS HERE! | << 1 2 3 4 5 ... 8 >>Post Follow-up |
| Kevin_in_GA 4,599 posts msg #93830 - Ignore Kevin_in_GA modified |
6/14/2010 1:42:44 PM I recently finished reading "High Probability ETF Trading" by Larry Connors. My general assessment is that there is real value to reading this for anyone who is looking for simple, effective, and profitable short-term trading systems. He presents 6 basic trading systems (actually 12, in that these can be traded both long and short) that are backtested against a set of 20 well-established and highly liquid ETFs. He provides examples of each type of trade and shows the success rates for both regular and "aggressive" trading strategies - here aggressive means doubling down on your bet if the trade moves against you and a second buy signal is generated. I thought it would be helpful to others here if I spent a little time reducing these systems to a set of tradable filters. In the next few posts I will spend time on each of the six systems Connors' uses in the book, providing filters for each system, and include Connors' backtest data (as well as my own for each filter going back to 1/1/2007). Note that in comparing these sets of results, Stockfetcher does not do its backtesting the same way as Connors does - he uses "buy/sell at the close of the day the signal is triggered" while SF uses "buy/sell on the open of the day following the signal being triggered." This actually helps us, in that the two methods of calculation provide for different exits and therefore will be reflective of just how robust the trading method is - if both methods work about the same, then I see it as saying that a good trade was entered and was profitable enough to exit at several different points that meet the trader's needs as well. As always, comments and critique are encouraged. Kevin |
| brank20 41 posts msg #93831 - Ignore brank20 |
6/14/2010 1:54:52 PM While stock fetcher uses the open of the following day by default, it can be changed. Simply change the "open" to "close 1 day ago". |
| Kevin_in_GA 4,599 posts msg #93832 - Ignore Kevin_in_GA modified |
6/14/2010 2:00:35 PM CONNORS FILTER #1: 3 DAY HIGH/LOW METHOD The 3 day high/low filter looks for ETFs that have made lower highs and lower lows for three consecutive days. Long Trade Rules: 1. ETF is above MA(200) 2. ETF is below MA(5) 3. ETF has made three consecutive lower lows 4. ETF has made three consecutive lower highs BUY on the close of the day these criteria are met. SELL on the close of the day the ETF closes above its MA(5). Connors' backtest results (709 trades from inception of the ETF through 12/31/08) show a win percentage of 76.9%, an average gain of 0.66% per trade, and an average holding time of 3.3 days. My backtesting data (258 trades from 12/31/2006 until today from the same set of ETFs) show a lower win percentage of 67.5%, an average gain of only 0.50% per trade, and the same average holding time of 3.3 days. Short Trade Rules: 1. ETF is below MA(200) 2. ETF is above MA(5) 3. ETF has made three consecutive higher lows 4. ETF has made three consecutive higher highs SHORT on the close of the day these criteria are met. COVER on the close of the day the ETF closes below its MA(5). Connors' backtest results (400 trades from inception of the ETF through 12/31/08) show a win percentage of 71.5%, an average gain of 0.88% per trade, and an average holding time of 3.6 days. My backtesting data (184 trades from 12/31/2006 until today from the same set of ETFs) show a lower win percentage of 68.5%, a higher average gain of 1.26% per trade, and an average holding time of 3.25 days. COMMENTS: The short version of this filter seems to be the better play, but both filters work well and can be traded profitably. |
| Kevin_in_GA 4,599 posts msg #93835 - Ignore Kevin_in_GA |
6/14/2010 2:09:29 PM CONNORS FILTER #2: RSI(4) METHOD The RSI(4) filter looks for ETFs that have closed below 25 (for longs) or above 75 (for shorts). Long Trade Rules: 1. ETF is above MA(200) 2. RSI(4) below 25 at the close BUY on the close of the day these criteria are met. SELL on the close of the day the RSI(4) closes back above 55. Connors' backtest results (786 trades from inception of the ETF through 12/31/08) show a win percentage of 76.7%, an average gain of 1.06% per trade, and an average holding time of 6.2 days. My backtesting data (321 trades from 12/31/2006 until today from the same set of ETFs) show a lower win percentage of 78.5%, ahigher average gain of 1.28% per trade, and a slightly shorter average holding time of 5.7 days. Same comments regarding when the trade was closed - looks slightly better to hold until the open, but remember that these are not the same set of trades (different set of trades over a different timeframe). Short Trade Rules: 1. ETF is below MA(200) 2. RSI(4) is above 75 SHORT on the close of the day these criteria are met. COVER on the close of the day the RSI(4) close back below 45. Connors' backtest results (383 trades from inception of the ETF through 12/31/08) show a win percentage of 68.1%, an average gain of 1.26% per trade, and an average holding time of 7.4 days. My backtesting data (193 trades from 12/31/2006 until today from the same set of ETFs) show a higher win percentage of 74.1%, a higher average gain of 1.87% per trade, and a shorter average holding time of 6.3 days. COMMENTS: Both sides of this filter trade profitably. The more recent backtesting data actually shows a meaningful improvement over Connors’ historical backtests. |
| Kevin_in_GA 4,599 posts msg #93836 - Ignore Kevin_in_GA |
6/14/2010 2:09:57 PM CONNORS FILTER #3: R3 METHOD The R3 filter looks for ETFs where the RSI(2) has dropped/risen for three days in a row, and is under 10/above 90. Long Trade Rules: 1. ETF is above MA(200) 2. RSI(2) decreasing for 3 days 3. RSI(2) below 60 3 days ago 4. RSI(2) below 10 at the close BUY on the close of the day these criteria are met. SELL on the close of the day the RSI(2) closes back above 70. Connors' backtest results (700 trades from inception of the ETF through 12/31/08) show a win percentage of 75.9%, an average gain of 0.92% per trade, and an average holding time of 5.0 days. My backtesting data (177 trades from 12/31/2006 until today from the same set of ETFs) show a similar win percentage of 75.1%, a higher average gain of 1.34% per trade, and a shorter average holding time of 4.3 days. Also looks slightly better to hold until the open of the next day, but remember that these are not the same set of trades (different set of trades over a different timeframe). Short Trade Rules: 1. ETF is below MA(200) 2. RSI(2) increasing for 3 days 3. RSI(2) above 40 3 days ago 4. RSI(2) above 90 at close SHORT on the close of the day these criteria are met. COVER on the close of the day the RSI(2) close back below 30. Connors' backtest results (383 trades from inception of the ETF through 12/31/08) show a win percentage of 70.4%, an average gain of 1.15% per trade, and an average holding time of 5.2 days. My backtesting data (145 trades from 12/31/2006 until today from the same set of ETFs) show a slightly higher win percentage of 71.7%, a higher average gain of 1.48% per trade, and a shorter average holding time of 4.2 days. COMMENTS: Both sides of this filter trade profitably. The more recent backtesting data confirms Connors' earlier data, and in fact is slightly better on both sides of the trade. |
| Kevin_in_GA 4,599 posts msg #93837 - Ignore Kevin_in_GA |
6/14/2010 2:10:30 PM CONNORS FILTER #4: THE %B METHOD The %B filter looks for ETFs where the Bollinger %B(20,2) has been below 0.2/above 0.8 for 3 consecutive days. Long Trade Rules: 1. ETF is above MA(200) 2. Bollinger %B(20,2) below 0.2 for the past three days BUY on the close of the day these criteria are met. SELL on the close of the day the %B closes back above 0.8. Connors' backtest results (1,014 trades from inception of the ETF through 12/31/08) show a win percentage of 76.5%, an average gain of 0.70% per trade, and an average holding time of 4.2 days. My backtesting data (on only 138 trades from 12/31/2006 until today from the same set of ETFs) show a higher win percentage of 79.0%, a much higher average gain of 2.65% per trade, but also a much longer holding time of 16.6 days. NOTE: On this filter, every trade since 4/16/2010 has been negative. Prior to that point, the numbers were pretty impressive - and 87% win rate with an average gain of 3.58%. Short Trade Rules: 1. ETF is below MA(200) 2. Bollinger %B(20,2) above 0.8 for the past three days SHORT on the close of the day these criteria are met. COVER on the close of the day the %B close below 0.2. Connors' backtest results (606 trades from inception of the ETF through 12/31/08) show a win percentage of 70.1%, an average gain of 0.95% per trade, and an average holding time of 4.5 days. My backtesting data (107 trades from 12/31/2006 until today from the same set of ETFs) show a lower win percentage of 61.7%, a much lower average gain of only 0.04% per trade, and a MUCH longer average holding time of 26.0 days. COMMENTS: Much better playing this filter on the long trades rather than the short ones. I am surprised that the differences in win percentage and overall performance. |
| Kevin_in_GA 4,599 posts msg #93838 - Ignore Kevin_in_GA |
6/14/2010 2:11:05 PM CONNORS FILTER #5: THE MULTIPLE DAYS UP/MULTIPLE DAYS DOWN METHOD This filter looks for ETFs where the close has been up/down for 4 out of the last 5 days. Long Trade Rules: 1. ETF is above MA(200) 2. ETF is below MA(5) 2. Close has been lower for 4 out of the last 5 days BUY on the close of the day these criteria are met. SELL on the close of the day the ETF closes back above its MA(5) Connors' backtest results (1,071 trades from inception of the ETF through 12/31/08) show a win percentage of 73.6%, an average gain of 0.50% per trade, and an average holding time of 3.3 days. My backtesting data (on only 415 trades from 12/31/2006 until today from the same set of ETFs) show a lower win percentage of 68.5%, a similar average gain of 0.49% per trade, and a similar holding time of 3.2 days. Short Trade Rules: 1. ETF is below MA(200) 2. ETF is above MA(5) 3. close has been higher for 4 out of the last 5 days SHORT on the close of the day these criteria are met. COVER on the close of the day the ETF closes back below its MA(5) Connors' backtest results (606 trades from inception of the ETF through 12/31/08) show a win percentage of 71.1%, an average gain of 0.80% per trade, and an average holding time of 3.4 days. My backtesting data (306 trades from 12/31/2006 until today from the same set of ETFs) show a slightly higher win percentage of 72.5%, a higher average gain of 1.08% per trade, and an average holding time of 3.5 days. COMMENTS: Better playing this filter on the short trades rather than the long ones. |
| Kevin_in_GA 4,599 posts msg #93840 - Ignore Kevin_in_GA |
6/14/2010 2:11:40 PM CONNORS FILTER #6: RSI(2) 10/6 (90/94) METHOD This filter looks for ETFs where the RSI(2) is at extremes (either below 6 or above 94). Long Trade Rules: 1. ETF is above MA(200) 2. RSI(2) below 10 1 day ago 2. RSI(2) below 6 BUY on the close of the day these criteria are met. SELL on the close of the day the ETF closes back above its MA(5) Connors' backtest results (1,075 trades from inception of the ETF through 12/31/08) show a win percentage of 81.90%, an average gain of 0.93% per trade, and an average holding time of 3.7 days. My backtesting data (on only 153 trades from 12/31/2006 until today from the same set of ETFs) show a lower win percentage of 73.9%, a lower average gain of 0.79% per trade, and a shorter average holding time of 3.3 days. Still a good performing filter – the differences seen here may be due to the backtest exit criteria discussed earlier. Short Trade Rules: 1. ETF is below MA(200) 2. RSI(2) above 90 1 day ago 3. RSI(2) above 94 SHORT on the close of the day these criteria are met. COVER on the close of the day the ETF closes back below its MA(5) Connors' backtest results (566 trades from inception of the ETF through 12/31/08) show a win percentage of 76.0%, an average gain of 1.56% per trade, and an average holding time of 3.8 days. My backtesting data (182 trades from 12/31/2006 until today from the same set of ETFs) show a much lower win percentage of only 59.0%, an average gain of only 0.84% per trade, and an average holding time of 2.0 days. COMMENTS: Much better playing this filter on the long trades rather than the short ones. |
| Kevin_in_GA 4,599 posts msg #93841 - Ignore Kevin_in_GA modified |
6/14/2010 2:17:49 PM BRINGING IT ALL TOGETHER If you're like me, you don't want to have to run 12 separate filters each day to see if any of these generate new trades. So let's make life a little simpler and combine all of the long filters in one mega-filter, and do the same for the short versions as well. CONNORS' COMPOSITE LONG FILTER You can see that what I have done is to make each individual filter entry criterion into a user variable that "counts" if that criterion is hit today, and if so assigns a value of "1" - if it is not hit, it is assigned a value of "0". Then you simply multiply all of the individual variables together and only if all are hit will it return a "1". If not, the filter will not return any stocks. Each column therefore is the output of a separate filter - what is nice here is that you can see when a given ETF is returned by multiple filters (hopefully improving its likelihood of a positive return). CONNORS' COMPOSITE SHORT FILTER Enjoy .... |
| jnafach 74 posts msg #93846 - Ignore jnafach |
6/14/2010 4:03:53 PM Hi Kevin, I want to thank you for all that you do, and being open to share the info, especially I am looking for ways to regain may fortune I lost that I could have buy house for it Anyway, I read this book last year I looked at doing it but I found it little hard to practice, the opther problem is that you will miss a lot of days where RSI raise >90 and stay there and loose a lot of gain potential, I have seen in one of your posts that you may try it, I hope you luck but the way is one of those is the TPS that they claim almost 90% success also if you never visited you can look at their power ratings is tradingmarkets.com as this is free charts with scores on it, I think mainly based on RSI(2) which may cut a lot on your gains especially in years like last 2-3 years thanks |
| StockFetcher Forums · Filter Exchange · \"HIGH PROBABILITY ETF TRADING\" BY LARRY CONNORS - GET YOUR FILTERS HERE! | << 1 2 3 4 5 ... 8 >>Post Follow-up |
Copyright 2022 - Vestyl Software L.L.C.•Terms of Service | License | Questions or comments? Contact Us
EOD Data sources: DDFPlus & CSI Data
Quotes delayed during active market hours. Delay times are at least 15 mins for NASDAQ, 20 mins for NYSE and Amex. Delayed intraday data provided by DDFPlus