Kevin_in_GA
4,599 posts
msg #101462
- Ignore Kevin_in_GA |
7/6/2011 6:00:28 PM
Why would you mess with a well-tested system just because it is telling you to stay in cash? Maybe it's right???
|
decipherlinda
133 posts
msg #101468
- Ignore decipherlinda |
7/7/2011 12:42:58 AM
Kevin, I don't have time to get into your analysis, etc., but John Bollinger in his book on Bollinger Bands said that actually stocks fell outside the bands 5% of the time when you set them to 1.8 instead of 2.0. I don't know if anyone else has mentioned that in this thread, or if it makes any significant difference to your work. Cheers!
|
jkinghome
7 posts
msg #101477
- Ignore jkinghome |
7/7/2011 1:01:42 PM
Hi Kevin,
I never got an exit signal for SJM but your performance showed as closed on 7/1. Even when I do a backtest, it shows as still open.
Thanks!
|
campbellb75
101 posts
msg #101481
- Ignore campbellb75 |
7/7/2011 1:36:08 PM
@jkinghome - Kevin's stats closed out all open positions to let you know how they were performing up to that point. SJM hasn't been triggered to close the position yet, so it should still be an open position if you're trading off of this filter.
HTS
|
knightvictor
7 posts
msg #101486
- Ignore knightvictor |
7/7/2011 6:18:28 PM
Kevin_in_GA
7/6/2011 3:02:01 PM
No, that is correct. There have not been any buys signaled since 6/24. That is not necessarily a bad thing. Probably reflects the ridiculous move up last week made by the S&P. The system saw this as a time to sell into strength.
----------------------------------------------------------------------------------------
Thank you very much for the follow up.
-Dan
|
mahkoh
1,065 posts
msg #101502
- Ignore mahkoh |
7/8/2011 2:48:42 PM
Kevin_in_GA
1,926 posts
msg #101462
- Ignore Kevin_in_GA 7/6/2011 6:00:28 PM
Why would you mess with a well-tested system just because it is telling you to stay in cash? Maybe it's right???
Quote from your opening post:
"This still returned too many stocks, and during market corrections the system took some big hits. So I simply added a criteria that included “close above MA(XX)” for each stock, where XX was 50, 100, 150 or 200 days. This hopefully keeps you out of taking long positions on stocks that are tanking (keeps you “buying the dips”).
This definitely helped, and the system dramatically beat out the ^SPX, but the equity curves were still choppy and the system called for a lot of trades. "
Apparently your choice for a close above the MA syntax was not the part of the filter that you used the best of your skills and talents on and I think it would be interesting to see what improvements can be made here. After all you were basically looking to weed out stocks that are in a downtrend.
|
Kevin_in_GA
4,599 posts
msg #101504
- Ignore Kevin_in_GA |
7/8/2011 3:00:52 PM
Yes and no. The core concept is to identify non-equilibrium relationships between S&P 500 stocks and the index. Building from that, one tries to improve profitability and percent wins by adding intiutive system components, but always making sure that they are part of an optimization and rigorous backtest.
The equity curve shown in my first post contains the close above MA(200) component. Removing it resulted in poorer performance. That is why I made the comment about changing a system that has done well historically just because it is not currently signaling any trades.
Change this as you see fit, but don't expect the same performance going forward.
|
mahkoh
1,065 posts
msg #101512
- Ignore mahkoh |
7/8/2011 5:53:29 PM
Kevin, you did indeed make it clear in your opening post that adding the 200 MA improved results.
I do however believe that there may be possibilities for further enhancement at this point. Using a syntax like "x day slope of the close is above 0" may produce results that address the requirement of being in an uptrend while being below the 200 MA.
I use your filter together with weekly and monthly pivot point analysis and have found that a reversal appears both more likely and more powerful when the filter's results are close to these support levels.
|
Kevin_in_GA
4,599 posts
msg #101514
- Ignore Kevin_in_GA |
7/8/2011 5:58:31 PM
Sounds interesting. I posted this not as a static result but one that could be evolved to be better, and to stimulate discussion on how good filters should be designed.
If you have modified it to improve overall profit or win%, please post.
|
duncanhoo
14 posts
msg #101557
- Ignore duncanhoo |
7/11/2011 9:21:59 AM
Kevin - thanks for all your hard work. just reading your posts/reasoning etc. on back testing has taught me tons. Just read through this whole thread and wondering if anyone is trading this with real $?
|