einok
msg #93933
- Ignore einok |
6/16/2010 11:19:23 PM
Bartune1,
You wrote "here is also a VIX variance signal today - three days stretched > 10% below the 10 dma"
I found a rule that the VIX be > 10% below the 10 day ma and the close be above the open (the idea being that momentum has shifted towards the mean)
Same with short VIX > 10% above the 10 day ma and the close be below the open.
|
petee
2 posts
msg #94002
- Ignore petee |
6/18/2010 6:57:13 AM
Thanks for sharing O'connors filters. I did notice one discrpancy between the filters listed and the ones O'Connor lists in his book.
The %b filter states "The %B filter looks for ETFs where the Bollinger %B(20,2) has been below 0.2/above 0.8 for 3 consecutive days." - this filter might work fine. However, O'Conner uses a %b(5,1) filter rather than the standard %b(20,2) filter.
|
Kevin_in_GA
4,599 posts
msg #94086
- Ignore Kevin_in_GA
modified |
6/20/2010 12:47:51 PM
The %b filter states "The %B filter looks for ETFs where the Bollinger %B(20,2) has been below 0.2/above 0.8 for 3 consecutive days." - this filter might work fine. However, O'Conner uses a %b(5,1) filter rather than the standard %b(20,2) filter.
++++++++
I re-read the entire chapter on %B, and there is no mention or reference to the settings you mention. I therefore assume that he is using the standard Bollinger Band settings of (20,2).
A similar backtest on your proposed settings for the long filter yields the following results:
389 trades
71.7% win rate
0.77% average gain
4.2 day average hold time
As I wrote earlier on this filter with the standard settings:
"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%.
I think I will stick with the standard settings for now.
|
Kevin_in_GA
4,599 posts
msg #94153
- Ignore Kevin_in_GA |
6/22/2010 9:32:04 AM
Some results from each composite filter for today (6/22/2010)
Long Filter -
XLY - I also got this as oversold on my pair trade filter, since the XLY:SPY ratio is more than 2 standard deviations below its MA(13). Likely reversion to the MA.
Short Filter -
EWH - composite score of 4 (out of a possible 6). This has been on the screen since 6/10 when it was at 14.99. It has gone up 4.2% since then, so it might be time to look at this as a good opportunity to short.
EWZ - composite score of 4 (out of a possible 6). Same as above (since 6/10 when it was at 65.10 - up 4.3% since).
|
cwn6161
40 posts
msg #94160
- Ignore cwn6161 |
6/22/2010 10:37:08 AM
Stop losses are pretty obvious for these systems - you can keep it the same for all. How do you determine when to exit with a profit though, since the systems have different exits. Are you just picking one and going with it? A stock with 4/6 indis saying to short is a great thing - I'm just not sure which of the 4 to use to get out.
|
Kevin_in_GA
4,599 posts
msg #94162
- Ignore Kevin_in_GA |
6/22/2010 10:47:36 AM
These systems are designed to trade without stop losses - you should read Connors' work on this. He advocates doubling down if the trade moves against you, and his historical data indicates that this improves both win % and profit.
As to which exit to take - assuming that 4 separate filters call this one as a long or short, then take the first filter exit that is triggered. A simple default would be to use the cross above/below the MA(5), as this is core to many of Connors' filter designs.
|
cwn6161
40 posts
msg #94163
- Ignore cwn6161
modified |
6/22/2010 10:59:37 AM
Ah didn't realize this Kevin - I may check out his book.
|
Kevin_in_GA
4,599 posts
msg #94164
- Ignore Kevin_in_GA |
6/22/2010 11:16:00 AM
I would not spend the money on it right now - all of the core trading strategies I have summarized here, and provided you with the filters to trade them. Save the cash for something else.
|
wkloss
231 posts
msg #94221
- Ignore wkloss |
6/23/2010 4:06:28 PM
Kevin,
Have you backtested the composites? I tried using 12-31-07 to 12-31-09 with no profit or stop losses, no max # of days to hold etc. Results weren't good. In fairness, I recognize that certain aspects of Connors strategies cannot be coded.
Just wondered your thoughts on this?
Bill
|
df0000
2 posts
msg #99633
- Ignore df0000 |
3/12/2011 4:33:27 PM
There is significant discordance between the ROI that the backtest results show and the return on account, because you're sitting in cash most of the time, and the number of ETFs you're trading mandates smaller amounts for each trade. In my backtesting, an ROI of 70% translates into a return on account of about 6% over two years. You can boost your return on account considerably by increasing your trade size (reduce max number of daily or open trades).
|