what back testing results are considered worthy? - StockFetcher.com Stock Screener

StockFetcher Forums · General Discussion · what back testing results are considered worthy?

<< >>Post Follow-up

psheridan050
52 posts
msg #46870
- Ignore psheridan050

9/12/2006 10:03:52 AM

Hi all,

Im new to StockFetcher and im playing around with the backtesting feature. Im trying to get a feel for what types of results are worth paying attention to. Ive seen in other posts where people have amazing results, but other people just say its good results on paper and would fail in real life trading becasue of one attribute or another of the results. Ive attached one of my own results but am unsure what i should be really paying attention to when deciding if a filter is worthy. Is ROI the best ting to look at? Average draw down? Reward/Risk ratio? How high should the numbers be to be considered a prfitable tading filter? Thanks for your advice.

Approach Information

Approach Name: base5

Test started on 05/11/2006 ended on 09/11/2006, covering 84 days

Filter used:

End of the Trend (down) (saved filter)

Trade Statistics

There were 203 total stocks entered. Of those, 200 or 98.52% were complete and 3 or 1.48% were open.

Of the 200 completed trades, 108 trades or 54.00%resulted in a net gain.

Your average net change for completed trades was: 0.40%.

The average draw down of your approach was: -1.59%.

The average max profit of your approach was: 2.00%

The Reward/Risk ratio for this approach is: 1.54

Annualized Return on Investment (ROI): 101.79%, the ROI of ^SPX was: -5.13%.

Exit Statistics

Stop Loss was triggered 1 times or 0.50% of the time.

Stop Profit was triggered 0 times or 0.00% of the time.

Trailing Stop Loss was triggered 6 times or 3.00% of the time.

You held for the maximum period of time (1 days) 193 times or 96.50% of the time.

An exit trigger was executed 0 times or 0.00% of the time.

Statistics By Holding Period
	Completed	2 day chg	3 day chg	4 day chg	5 day chg	10 day chg
Winners:	108	92	83	77	71	77
Losers:	85	103	115	118	119	91
Win/Loss Ratio:	1.27:1	0.89:1	0.72:1	0.65:1	0.60:1	0.85:1
Net Change:	0.40%	-0.83%	-1.59%	-2.10%	-2.29%	-2.20%

Statistics By Variable: Match Price
	<6	<9	<12	<15	<18	<21	<24	<27	<30	<33
Completed	1:2	3:6	4:4	2:1	7:3	11:18	16:8	23:14	26:21	15:8
2 day chg	0:3	3:9	4:4	0:3	3:6	18:10	9:14	19:18	23:26	13:10
3 day chg	0:3	2:10	3:5	0:3	5:5	18:11	6:18	16:19	24:27	9:14
4 day chg	0:3	1:11	3:5	0:2	4:6	16:13	4:19	15:20	23:28	11:11
5 day chg	0:3	2:10	1:6	0:1	6:4	15:14	5:16	12:22	21:30	9:13
10 day chg	0:3	4:8	0:6	-	4:6	15:10	10:11	13:11	24:21	7:15

Statistics By Variable: Average Volume
	<2.5M	<5.0M	<7.5M	<10.0M	<12.5M	<15.0M	<17.5M	<20.0M	<22.5M	<25.0M
Completed	73:50	22:17	3:0	-	3:5	1:1	1:1	4:10	1:1	-
2 day chg	56:69	15:26	2:1	-	6:2	1:1	2:0	8:4	2:0	-
3 day chg	48:77	17:25	0:3	-	4:4	1:1	2:0	10:4	1:1	-
4 day chg	43:80	16:25	0:3	-	4:4	1:1	2:0	10:4	1:1	-
5 day chg	38:80	17:24	0:3	-	3:5	1:1	2:0	9:5	1:1	-
10 day chg	36:64	21:18	1:2	-	4:2	1:1	2:0	10:4	2:0	-

traderblues
195 posts
msg #46873
- Ignore traderblues

9/12/2006 10:40:33 AM

First of all, people on this discussion can contribute a wealth of information and help you along the way. But just because someone says something will not work doesn't make it so. Be true to your own intuition, study charts, practice first by paper trading, then trade real money a little bit at a time. Then adjust and try again.

Here's what I look at:

1. Average gain per trade. For your backtest, your average gain per trade is only 0.4%. Keep in mind that this does not include commissions and slippage. By the time you factor these in, your gain per trade might be gone.

2. Reward/Risk ratio. If it's not above 2, I don't even consider trading it. I try to shoot for 3 or better.

3. How many trades. 200 trades for a 6 month period comes out to about 8 trades per week. Will 8 trades per week at 0.4% per trade(less commissions & slippage) satisfy your goal? Work it out on paper.

4. Drawdowns. I've noticed that the 'equity' chart that is included in the backtest results does not account for drawdowns on open trades. It only keeps track of winning vs. losing trades. So a chart that shows low drawdowns does not show the whole picture. The average drawdown per trade stat gives a little better picture, but there is nothing that shows maximum drawdown.

5. Is it tradeable? If it's a daytrading system, are you going to be able to watch your trades during the day? Do you need to? Can you easily buy and sell the stocks that the system picks?

psheridan050
52 posts
msg #46874
- Ignore psheridan050

9/12/2006 11:09:44 AM

my plans for trading are to basically buy around 10am and sell end of day. so no day trading. i have a margined account with a tad over 200k buying power for my trading. i set the filter to trade 5 stocks at most at one time. i would purchase 1000 shares at a time so i could load up on all the picks at one time if needed. commision is 8$ per trade and slippage is probably at 10$ a trade or so. i have a 12% stoploss for this particular filter. shouldnt that remove any huge drawdowns? as for tradability of the picks i use Average Volume(30) is above 1000000 to ensure i can get in and out fast nuff. i appreciate the comments. i would like to hear other ideas too

__fetcheruser123
msg #46882
- Ignore __fetcheruser123

9/12/2006 12:52:23 PM

Lets say you're only making one trade with all your $200k of buying power (just for ease of calculation).

Average gain:
$200,000 * 0.40% = $800.

Maximum drawdown:
$200,000 * -12% = -$24,000.

So, basically, with each trade you're only making about $800 for every $24k you risk.

Now, say you have a string of bad luck trades, resulting in the full $24k loss each time. I'm guessing you have margin of about 4/1... So, maybe $50k cash in your account? It would only take a couple hard losses before you wiped out your account completely... Those are really bad odds, especially when there are other filters/systems with much better odds.

My advice is to find a better system and learn about risk/reward and position sizing before you start trading. The book Trade Your Way To Financial freedom is a great book on the subject. Don't let the title throw you off. ;)

__fetcheruser123
msg #46883
- Ignore __fetcheruser123

9/12/2006 12:58:17 PM

Also, your backtest is only for four months... That is a really small timeframe for a backtest and could just represent an anomaly in the market.

traderblues
195 posts
msg #46887
- Ignore traderblues

9/12/2006 1:12:33 PM

I agree about the backtest period. Do yourself a favor and buy 3 months of the advanced subscription. If you're looking to trade 200k, what's $44.95? Having a better system will return much more than that. IMHO this is one of the best deals out there.

psheridan050
52 posts
msg #46888
- Ignore psheridan050

9/12/2006 1:14:15 PM

this information helps out alot. im sure there are some pretty good stock filters on this website, but does anyone have a favorite that is pretty representative of what to look for? if not the actual filter itself then the results. i only backtested 4 months becuase i dont have the advanced version of SF. to be honest, when i was looking t the drawdown, i was not considering it agasint my whole account, but i guess the example you used makes sense as well. my guess is that i would have to make a filter that has a stop loss of around 2% in order to keep risk to a resonable level. any other comments out there on what has been discussed here?

psheridan050
52 posts
msg #46889
- Ignore psheridan050

9/12/2006 1:23:04 PM

i use tradestation for trading. within that environment i develop strategies and such. the problem with tradestaton is that it is a pain to backtest a strategy on multiple stocks. which is why im using SF. i also use SF to gather ideas and test things out fast. the example filter i posted here was just something i tossed together. my point in posting my questions is to try to get a good idea on how filters are rated and which critera should i be trying to max out. at first i thought it was ROI. but the responses i have recieved oint otherwise. also, as for my backtesting of 4 months, i can always go in there and change the dates for another 4 momthsof backtesting ending where the first backtestinf started. i just cant do more than 4 months at a time.

StockFetcher Forums · General Discussion · what back testing results are considered worthy?

<< >>Post Follow-up