springhill
74 posts
msg #109070
- Ignore springhill |
11/29/2012 4:46:05 PM
Hello Kevin,
I have been manually back testing the code above.
Some picks are very good but going back 10 days or more the filter starts to pick stocks that are definitely in a downtrend.
Offset the above code to ..
offset is 15
and look at most of the 92 picks that continue down significantly for some time.
I see that you use a "trigger" in the back testing you do that has to cross above the low of the pick day.
Does that same rule apply in manually executing the trade?
Could an entry be calculated that adds .02 cents to the trigger and be a valid entry criteria?
set {entry, trigger + .02}
add column entry.
Thanks for all your sharing of stuff with us folks here on SF !
|
Kevin_in_GA
4,599 posts
msg #109071
- Ignore Kevin_in_GA
modified |
11/29/2012 6:10:23 PM
The entry is a limit entry for a 6% pullback from today's close (that is after today also showing a 6% pullback from yesterday's close).
Example - stock price closed yesterday at $100. Today's low was below $94, and the stock closed somewhere in the bottom 20% of today's range. For simplicity let's say that range was from $100 to $93.99 and the stock closed at $95.
You would put in a limit order at $95 * 0.94 or $89.30. The low tomorrow must cross below $89.30 (another 6% pullback) to trigger an entry.
Is that consistent with your entry strategy? If not that might explain your results.
Also, the general trend is not as relevant as the price at the exit trigger - a few days is the average hold for these trades.
|
bjfeen
24 posts
msg #109084
- Ignore bjfeen
modified |
11/30/2012 4:08:26 PM
Kevin, curious... is there a stop loss used in your backtest?
For instance, the 2nd filter in your post picked UGAZ on Thursday. If I understand correctly, we would have entered today @ $28.03 (Thursday's close of $29.82 * 0.94), but UGAZ continued to drop beyond $28.03.
EDIT: Just checked the closing price and UGAZ came off it's lows to close at $28.02, almost exactly 6%. LOL
|
mahkoh
1,065 posts
msg #109101
- Ignore mahkoh |
12/3/2012 6:21:05 AM
Kevin, you state that a trade is triggered when the stock pulls back 6 % after having experienced a similar pullback the day before. The SF code however only requires one pullback?
|
Kevin_in_GA
4,599 posts
msg #109102
- Ignore Kevin_in_GA |
12/3/2012 8:29:42 AM
Kevin, you state that a trade is triggered when the stock pulls back 6 % after having experienced a similar pullback the day before. The SF code however only requires one pullback?
Right. It generates a candidate list for the next day. You would put limit orders in at 6% below the prior day's close and wait to see which (if any) trigger.
|
Kevin_in_GA
4,599 posts
msg #109113
- Ignore Kevin_in_GA |
12/4/2012 11:19:40 AM
Here's a quick update. The candidate list generated after the close yesterday was as follows:
ACTV
FIVE
HGT
NFLX
SKUL
SSYS
You set a limit order at 6% below the day's close and wait.
As of 11 AM this morning SSYS triggered (at 65.85) and is currently at 67.36 for a gain of 2.3%
|
springhill
74 posts
msg #109117
- Ignore springhill |
12/4/2012 8:09:35 PM
Hi Kevin,
I modified the filter you posted so the selection was moved back to use close 2 days ago and low 1 day ago.
Then I test to see if the current day triggers below the close of 1 day ago
If the test is positive I set a variable and add a column "long"
Sort the list to put the "longs" on top so one can quickly look to see the results of a purchase on the "long" day.
I added columns with all the data used in making the selection.
Running the filter today delivers the same list of six you posted above for yesterday and puts SSYS to the top of the list.
When running offset back a day at a time, there are quite a few "longs" that do not move in the right direction.
Can you think of any data one could use to confirm going "long"
Thanks for all the things you share with us.
|
Kevin_in_GA
4,599 posts
msg #109149
- Ignore Kevin_in_GA
modified |
12/6/2012 10:37:35 AM
Entry yesterday in EDU at 17.00, which closed at 16.41. However, today it is currently trading at 17.63 (up 3.7%). RSI(2) at 46, so a sell will come soon. Might want to simply take the money off the table now and relax.
Previous entry in SSYS on 12/4 at 65.85 now really bearing fruit - currently trading at 69.62 (up 5.7%). SSYS RSI(2) now at 53, so it it probably a good time to sell and move on.
EDIT: 10:39 am - EDU now at 17.90. Exit city for a gain of 5.3% in two days. Thank you Larry Connors.
|
bjfeen
24 posts
msg #109154
- Ignore bjfeen
modified |
12/6/2012 12:10:12 PM
Not to mention that UGAZ in my post above crossed RSI(2) @ 50 yesterday and closed at 30.32... a gain of over 8% on initial entry at 28.03 last week Thursday. Nice!
|
BarTune1
441 posts
msg #109170
- Ignore BarTune1 |
12/6/2012 8:10:38 PM
In / out of 500 shares of EDU at 17.00 / 17.72 for a 4.24% gain. Just over 4% after commissions.
|