push5280
201 posts
msg #81687
- Ignore push5280 |
10/24/2009 2:24:37 PM
Here is a filter I came up with seems to have some good short term (1-20 days) potential. Play around with it. All tweaks are welcomed.
|
eddyadiel
18 posts
msg #81718
- Ignore eddyadiel
modified |
10/25/2009 11:04:27 AM
Is this filter for Long or Short???? 'Coz per my analysis...it has experienced a strong downtrend...don't know if it is potential for uptrending for the next 1-20 days as you said for the holding period....
|
push5280
201 posts
msg #81726
- Ignore push5280 |
10/25/2009 5:00:42 PM
this is a long swing filter. I've pulled all of the hits from 1/1/09 through 10/23/09 (930 to be exact) and I am putting together 14 day and 30 day numbers. Just something else I saw when I was pulling the hits, the ytd avg return on these 930 stocks, from the day they hit through 10/23 is 289.84% not too shabby.
|
trendscanner
265 posts
msg #81733
- Ignore trendscanner |
10/25/2009 7:28:34 PM
Push, I backtested the filter from 1/1/09 through last Friday, using the default criteria (10% stop loss, 10% profit stop, 25 day hold) and got the following results:
There were 729 total stocks entered. Of those, 706 or 96.84% were complete and 23 or 3.16% were open.
Of the 706 completed trades, 336 trades or 47.59%resulted in a net gain.
Your average net change for completed trades was: -0.13%.
The average draw down of your approach was: -9.75%.
The average max profit of your approach was: 10.79%
The Reward/Risk ratio for this approach is: 0.97
Annualized Return on Investment (ROI): -9.16%, the ROI of ^SPX was: 25.57%.
Exit Statistics
Stop Loss was triggered 356 times or 50.42% of the time.
Stop Profit was triggered 322 times or 45.61% of the time.
Trailing Stop Loss was triggered 0 times or 0.00% of the time.
You held for the maximum period of time (25 days) 28 times or 3.97% of the time.
An exit trigger was executed 0 times or 0.00% of the time.
Overall, the system underperformed the SPX by quite a bit.
If you conducted a backtest where you simply held those stock since they were bought and haven't sold them yet (e.g., buy and hold) I can understand that your backtest statistics probably look better than the ones above since the whole market has risen significantly since March.
What you have developed is a bottom fishing filter of sorts - one that finds a condition where a stochastic oscillator is oversold.But oscillators can remain oversold (or overbought) for an extended period of time. A stock's price doesn't have to reverse immediately simply because an oscillator has reached an oversold or overbought condition. Price can keep going up or down.
You might want to consider adding some qualifying conditions that suggest when a reversal is closer to occurring or has been confirmed. There are a number of ideas on this in old forum posts - things like close is above the open, close is above ema(5), slow stochastic has crossed above 10 or 20, etc.
|
push5280
201 posts
msg #81767
- Ignore push5280 |
10/26/2009 1:09:46 PM
Trend,
I've started going through these manually and I've made it through 6/23 - 9/25. I'm using the close the day of the hit and the highest high for 14 calendar days after the hit. from 6/23 - 9/25 there were 130 hits, one was a loss, two were a wash and the rest were up. The average is 32.05%.
|
trendscanner
265 posts
msg #81771
- Ignore trendscanner |
10/26/2009 2:31:45 PM
Push, I looked through the results of the backtest and you're right, for the time period you're looking at (late June to late sept) your filter had very good results.
I'd be interested in seeing your final conclusions for the filter after you work through the rest of the individual stocks
regards and good luck trading this week
|
push5280
201 posts
msg #81773
- Ignore push5280 |
10/26/2009 3:18:02 PM
I'm sure it's going to take some time to go through 900 hits, but I'll post a progress report every day.
|
push5280
201 posts
msg #81912
- Ignore push5280 |
10/27/2009 11:03:48 PM
ok, so I'm going through these manually to get a baseline. So far I have looked at every hit for the period of 3/9/2009 through 9/25/2009. Theoretically the entry would be the close of the day of the hit, exit is the high of the next fourteen calender days.
So far I have looked at 302 hits
1 stock was a loss .33%
9 stocks were a wash 2.98%
and 292 were winners 96.69%
The avg. return is 42.39% and the avg time held is 8 calendar days! So far so good!!!!!!!!
|
push5280
201 posts
msg #82011
- Ignore push5280 |
10/29/2009 10:54:15 AM
so as I said in an earlier post I am going through all of the results from 1/1/09 to current to get the true effect of this scan. I am using the close on the day it hits as the entry and the high of the next 14 calendar days as the exit. So far .......
3/2/09 - 9/25/09 529 hits
4 trades were a loss (.75%)
9 trades were a wash (1.70%)
and 516 were profitable (97.54%)
the avg. return on the 529 trades is 38.56% and an avg of 9 calendar days held!!!
Pretty good results so far!!
|
crunkle
54 posts
msg #82020
- Ignore crunkle |
10/29/2009 12:11:18 PM
Push,
Understand that there's many filters that work well in a rising market (your scan begins 3/02/09, right at a multi-year bottom) and don't work well at all in a declining one.
Before you begin to put real money out there, test your filter in the Jan/Feb 09 period to see what it does in a collapsing market or at least during a period where the market fluctuates in both directions, like the year 2007.
Good luck
|