Friday, May 22, 2009

Friday 05/22/09 - Stop-Loss Experiment

I went to bed around 2:30 AM last night, and slept in till 9:30 AM this morning. I got to the screens around 10:30 AM (central), and saw that price had bounced off 882, an area that I marked as support yesterday. I utilized just the 5-minute chart for trade placement today, and my focus was to experiment with my stop-loss management. I decided I would let the trade run and move my stop-loss to break-even once price had moved 2 points in my favor. My initial profit target was 4 points (1:2 R/R). I had a total of 3 trades.

I entered Long around 10:55 AM at 890.50 with a 2-pt stop and an initial 4-pt profit target. Once price moved 2 pts in my favor, I moved my stop-loss to break-even. Price then went 3 pts in my favor, at which point, it looked like I should exit the trade but I had to stick with the 4 pt target for this experiment's sake, and was exited at break-even around 11:25 AM (central).

Trade #2
I entered Long around 11:30 AM at 890, again with a 2 pt stop and 4 pt profit target. Once price moved 2 pts in favor, I moved the stop-loss to break-even. Price then continued and was +3.75 pts in my favor, a mere 1 tick away from my exit order. The 12:10 PM (central) candlestick signaled a reversal and I could have bailed for +3 points, but I stuck with the experiment and once again was stopped out at break-even at 12:35 PM.

Trade #3
I entered Long around 1:00 PM at 890.50, with a 2-pt stop and 4 pt profit target. I actually had to leave the house at this time, so put in the bracket orders for stop and profit, and left the PC. On this trade, price finally hit my profit target for +4 pts at 2:15 PM (central). The 2:20 PM candle would have signalled a short entry for me, but I wasn't at my desk.

ES (5-Min) - Trade Chart

I realize one day's price action doesn't say much, but it looks like I need to give my trades some room in the beginning. Moving stop-loss to break-even at +2 pts sounds reasonable from a risk management perspective. But then if I see a reversal signal, I also need to go ahead and exit at a profit. I easily left 3 pts on the table on my first two trades, so this could actually have been a +10 pt day in just 3 trades if I used discretion and exited the first two trades at a profit. Here's the Stats break-down:

# of Trades 3
Longs 3
Shorts 0
% Break-Even: 66.67
% Win 33.33
% Loss 0.00
Avg Win 4.00
Avg Loss 0.00
Largest Win (pts) 4.00
Largest Loss (pts) 0.00
Total Win (pts) 4.00
Total Loss (pts) 0.00
Net Gain/Loss (pts) 4.00
ES Daily Range 13.50
P/L as % of Daily Range 29.63

We have an open gap at 899.50, and at this point, the 876-882 area has proved to be support. Lets see if that holds next week. Have a fun and safe memorial day weekend!

ES (Daily) - Zoomed In. Price made a lower high and higher low today.


  1. I stumbled across your site by accident. I like your idea of using a journal. I left you an anonymous post earlier today.
    Since you have invested your time, I thought in return, I would leave some comments.

    With regards to your trades:
    a. I made the same trades as your #1 and #3. I have a rule that I do not trade between 11:30 and 1:30 pm EST. To much fluctuation and lower volume during this period. Since I have eliminated this time period my success rate has gone up by 7%.

    b. Your fixed 4-point target range is a bit challenging with your stop at b/e after 2 points. Normal market retracement is 1.75 which leaves you with little room on a day like today (an Inside Range day). I would not consider using this approach with any inside day let alone on a day with lower-than-average-volume due to traders taking a long holiday weekend.
    c. If you trade more than one contract, you are better off taking 1/2 of your position off at 2 points. Move your remaining to -2 stop loss. Then use s/r levels as an exit for 1/2 of your remaining. Once the price moves 4 points above entry, then move last 1/2 to b/e and hope to catch a market trend (which only happens 30% of the time). This approach requires a min. of 4 contracts, but works very well.

    Also, one of your trades stopped out and looked to get an immediate reversal as the price hit a fib +.618 support level and continued back up. If you are trading with less than one contract, instead of setting your s/l to b/e, I would instead place my stop-loss 1-2 ticks below the fib-levels or a reliable moving average (i.e. 50 ema on a 1 min chart). Yes, your risk is greater, with this approach, however, getting stopped out as the market continues in your direction is very deflating.

    d. Understanding market day-types.
    As I mentioned, today was an inside day. As such when the opening sell-off did not break below yesterday's low, normally the price would bounce back to the hod or overnight high. Then typically, on the second attempt at this extreme, the market will sell off and with little support, will drop back to the lod. Which is exactly what today's price action did. Many bloggers/traders were wondering what news caused the massive 3:30 pm sell-off today? The answer, there wasn't one, just profit takers and a lack of technical support. This happens regularly with inside days. Knowing this, I placed my long entries at 14:00 with a stop-limit set for the previous hod. Once it hit, I then reversed (like many did in the market).

    I hope that this info helps.


  2. Thanks for taking the time to post the detailed comment, Joseph! You're right; a static 2 pt stop/4 pt target is not the way to go.

  3. My algorithm also employs a volatility stop (7 bars 2.4 multiplier) which I have plotted on my chart. This is also a good level for setting your stop losses.

    Your s/l experiment could work if you are entering into a trending market. One thing I learned is that there is never one global right or wrong approach with these markets. It is knowing when to employ each technique.

    I am sure that you have experienced this during the long hours of back-testing and "tweaking" that we all must endure. Speaking of which, are you using standard indicators or have you customized your code? Which software platform do you use? Does your platform allow you to compile and run your algorithm to generate a statistical report of previous days?



At the minimum, please provide your name or Twitter handle when posting comments. Do not post as Anonymous. Comments that contain links to commercial websites will be marked as spam.