TrendCatcher 2.0
By
Rennie on Tuesday, January 27th, 2009 at 9:32 pm
One of the more frequent questions I receive is why an intraday TrendCatcher buy or sell signal was taken when I had just discussed the opposite implication in my commentary the night before. My answer has always been that TrendCatcher was a separate, purely mechanical strategy and that I didn’t want to interfere by imposing my own opinion. Upon review, it’s fairly clear that was the wrong approach. We’ve recently been going back over the intraday TrendCatcher signals and filtering them through the research discussed in the previous night’s column. Only short-term studies from the nightly column were considered – those lasting anywhere from 1-5 days. When there was a clear bias, which was usually the case, a TrendCatcher signal would only be taken the next day if it was in agreement with the short-term bias. Otherwise the signal was disregarded. While I knew recent performance would be enhanced by this filter, I was surprised at the effectiveness. The table below highlights every TrendCatcher signal since the middle of last year that did not disagree with the previous night’s short-term bias (listed at the end of the column).
Date Time Signal Entry Exit P/L Bias
07/01/08 12:00 SELL 1270.75 1284.50 -13.75 bearish
07/02/08 12:28 SELL 1277.50 1283.50 -6.00 bearish
07/07/08 12:30 SELL 1252.00 1263.00 -11.00 bearish
07/16/08 13:15 BUY 1231.00 1245.36 +14.36 bullish
07/24/08 13:58 SELL 1264.50 1252.54 +11.96 bearish
07/28/08 14:58 SELL 1239.00 1234.37 +4.63 bearish
07/29/08 11:46 BUY 1252.00 1263.19 +11.19 bullish +11.39
08/07/08 15:18 SELL 1270.00 1266.07 +3.93 bearish
08/08/08 10:23 BUY 1284.00 1296.32 +12.32 bullish
08/25/08 12:16 SELL 1268.00 1266.84 +1.16 bearish
08/27/08 12:45 BUY 1284.00 1281.66 -2.34 bullish +15.07
09/09/08 10:49 SELL 1253.50 1224.56 +28.94 bearish
09/17/08 10:36 SELL 1185.00 1156.39 +28.61 bearish
09/22/08 14:06 SELL 1225.00 1207.09 +17.91 mixed
09/25/08 13:02 BUY 1218.50 1209.18 -9.32 bullish
09/26/08 15:06 SELL 1195.00 1213.15 -18.15 mixed
09/29/08 10:22 SELL 1165.00 1106.39 +58.61 bearish
09/30/08 13:31 BUY 1152.75 1164.74 +11.99 bullish +118.59
10/03/08 11:01 BUY 1148.00 1130.20 -17.80(*) bullish
10/10/08 11:09 SELL 874.00 887.10 -13.10(*) bearish
10/21/08 12:00 SELL 958.00 972.50 -14.50(*) bearish
10/24/08 14:06 SELL 860.00 873.30 -13.30(*) bearish
10/27/08 14:40 SELL 872.00 848.92 +23.08 mixed
10/28/08 14:45 BUY 889.00 940.51 +51.51 mixed
10/30/08 13:07 BUY 941.00 954.09 +13.09 bullish +28.98
11/05/08 12:41 SELL 972.00 952.77 +19.23 bearish
11/06/08 10:36 SELL 930.00 904.88 +25.12 mixed
11/07/08 12:03 BUY 925.00 918.50 -6.50 bullish
11/10/08 12:46 SELL 918.50 926.00 -7.50 bearish
11/11/08 10:30 SELL 893.50 908.00 -14.50 bearish
11/12/08 10:15 SELL 882.00 852.30 +29.70 bearish
11/14/08 10:38 SELL 885.00 899.50 -14.50 bearish
11/19/08 10:25 SELL 856.00 806.58 +49.42 bearish
11/21/08 11:05 SELL 742.00 753.50 -11.50 bearish +68.97
12/04/08 15:12 SELL 848.00 845.22 +2.78 bearish
12/09/08 14:11 SELL 891.00 899.00 -8.00 bearish
12/11/08 14:57 SELL 882.00 873.59 +8.41 bearish
12/15/08 10:48 SELL 867.50 868.57 -1.07 bearish
12/16/08 14:59 BUY 897.00 913.18 +16.18 bullish
12/22/08 14:26 SELL 864.50 871.63 -7.13 bearish +11.17
01/07/09 14:23 SELL 907.00 906.65 +0.35 bearish
01/14/09 10:14 SELL 848.50 842.62 +5.88 bearish
01/21/09 15:16 BUY 831.00 840.24 +9.24 bullish +15.47
(*) = based on secondary price stop implemented in November '08
The filtered TrendCatcher strategy (currently dubbed TrendCatcher 2.0) triggered a total of 43 signals since July of last year, with no losing months and a theoretical profit of 260 S&P points. That’s significantly better than the standard TrendCatcher strategy (see record), which had twice as many trades in the same time period (89), posted three losing months (assuming January ends with a loss) and only made about the same amount in terms of theoretical S&P points (274). Average profit was 1.8% vs an average loss of 1.1% for the 2.0 strategy, nearly identical to the standard system’s average, and the percentage of winning trades increased from 54% to 58%. While we’re still wrapping up backtesting from the last three years, preliminary testing clearly suggests that filtering TrendCatcher signals through our short-term research greatly enhances the strategy’s overall performance.
TrendCatcher 2.0
By Rennie on Tuesday, January 27th, 2009 at 9:32 pmOne of the more frequent questions I receive is why an intraday TrendCatcher buy or sell signal was taken when I had just discussed the opposite implication in my commentary the night before. My answer has always been that TrendCatcher was a separate, purely mechanical strategy and that I didn’t want to interfere by imposing my own opinion. Upon review, it’s fairly clear that was the wrong approach. We’ve recently been going back over the intraday TrendCatcher signals and filtering them through the research discussed in the previous night’s column. Only short-term studies from the nightly column were considered – those lasting anywhere from 1-5 days. When there was a clear bias, which was usually the case, a TrendCatcher signal would only be taken the next day if it was in agreement with the short-term bias. Otherwise the signal was disregarded. While I knew recent performance would be enhanced by this filter, I was surprised at the effectiveness. The table below highlights every TrendCatcher signal since the middle of last year that did not disagree with the previous night’s short-term bias (listed at the end of the column).
The filtered TrendCatcher strategy (currently dubbed TrendCatcher 2.0) triggered a total of 43 signals since July of last year, with no losing months and a theoretical profit of 260 S&P points. That’s significantly better than the standard TrendCatcher strategy (see record), which had twice as many trades in the same time period (89), posted three losing months (assuming January ends with a loss) and only made about the same amount in terms of theoretical S&P points (274). Average profit was 1.8% vs an average loss of 1.1% for the 2.0 strategy, nearly identical to the standard system’s average, and the percentage of winning trades increased from 54% to 58%. While we’re still wrapping up backtesting from the last three years, preliminary testing clearly suggests that filtering TrendCatcher signals through our short-term research greatly enhances the strategy’s overall performance.