Jul 152012
 

In my recent blog posts I asked the question if it’s even possible to be consistently profitable trading large size in ES.  The jury is still out on that one, but what I do believe is that the way I was trading was close but no cigar.  So I wanted to change that.

And as I previously wrote, when working full time here in Europe I will only be able to trade the last 90 minutes of ES.  I don’t like entering positions in the last 15 minutes so that really gives an entry window of 75 minutes.

Edit: The first version of this chart had an error in the data, I’ve since corrected it.

I made a graph of the cash session (I usually call it RTH) highs and lows per 30min time period:

 

Using NY time, for each time period with the end time shown (example 10 = 9:30am-10am), the number of times that period was the high or low is plotted.  This is very interesting.  I have previously heard that the H/L of the day is often in the first 30 minutes, even in the first 10-15 minutes.  What I found is that, since 2007, the first 30min period is the least most likely to be the high or low.

Often when I verify things people say, I often find out that they’re not true.  So I like to verify everything myself.  If anyone has data to confirm or contradict this, please let me know.  Edit: Thanks for the feedback, I found my error and updated the post.

So my entry window will be in the last 3 time periods, and mostly in the 15 & 15:30 periods.  Since the last period is most very often high or low, that means I want to guess which it will be and try to enter pullbacks in that direction.

I might not get a pullback to a level/zone, and so I have to be able to enter on smaller pullbacks.  And that means going for smaller wins.  Which brings to mind this comment from MM:

…have you kept the stats on MFE/MAE? You’d be surprised at the difference between what we think works and what actually does work based on historical evidence.

It may be possible that your entries are good enough to hold your stops for a 1:1 at a WR that makes it all worthwhile and with less stress.

So I started going over my trades.  I found that going for more than 4 pts often left me wishing I hadn’t.  And not scaling at less than 2:1 often had me wishing I had.  So now several pieces of the puzzle were starting to fit together:

  • Use the projected High or Low direction for a bias (if we’re at or near the high, bias is up, etc.)
  • Set aside some of the bigger picture stuff and focus on what’s going on right now.
  • Use order flow to enter on small pullbacks in the current rotation in the direction of my bias
  • Use very tight stops (1.5-2.0 pts max)
  • Take a quick scale, ideally around 1.5-2 pts.  It’s OK if this gives 1:1 R:R
  • Go for 3-4 on second contract.

So I made these changes to my trading.  I removed several higher timeframe charts from my workspace and simplified a few others.  I removed all bund charts since I won’t be able to trade that any more.  It was a big simplification for me because some of the charts I removed had my own proprietary indicators on them.  It was a big step for me to “let them go”.  (But not all of them, at least not yet).

I tested these changes last week.  To maximize my testing, I traded all day, both Globex & US RTH, even though soon I’ll only be trading that last 90 minutes.  Since I no longer have to rely on trading for income and in order to remove pscyhological issues, I traded on simulator but treated it as real money as much as possible.  I considered it an experiment and I didn’t judge myself or my trading.  I wasn’t trying to be profitable, I was just trying to execute my plan and see what would happen.

The results were nothing short of remarkable:

There was a 68% increase in the number of trades per day, due to me not waiting for my levels.  The win rate was slightly higher.  But the expectancy went from $3 to $60 and the profit factor was 2.73.  Wow!

One week is not enough time to say this is a success, but the results are very promising.   I have a few more days to test it this coming week and after that my trading will be in the 90 minute window.  So here’s my plan:

I need to make sure my results will be similar when only trading the 90 minute window.  So when I start work (later this week) I’ll track my results during the 90 minute window separately.  After several weeks of consistent results on simulator, I’ll trade with real money.  If the results continue to be good I’ll stay on real money and this will provide a nice supplement to my consulting income.  If the results are not good, I will return to simulator in order to find out if the issue is the trading plan, the market conditions, or psychological.

Once consistent for a month or two with 2 lots real money, start increasing size slowly (1 lot per month maybe).   I only expect to have 1 setup per day so this is going to go very slow.  Last week had 5 trades/day, so effectively trading only the 90 minute window will take 5 times as long to acquire the same amount of trades.  Not only that but there will be same days I probably won’t be able to trade (Friday night pizza night, days I must work late at the office, etc.).

Last week my expectancy was roughly 1 pt/trade.  So at 1 trade/day that makes a pt per day or $50/day.  Going to 4 contracts would give $100/day.  10 contracts would be $500/day and then it starts to get substantial.  And of course if I can do this consistently then I could go back to full time trading.  But this time I will not trade full time until consistently profitable.  A couple years ago I thought I was very close and going full time would accelerate the process.  I was a bit optimistic then and so I’m going to be very conservative now.

I’ll be reporting on the progress.  Good luck for the upcoming week.

 

 

 

  6 Responses to “Review of last week’s test of trading adjustments”

  1. Very interesting results. What do you think was the biggest factor in the increase in profitability?

  2. I think it’s a combination of everything I listed. not sure it was any one specific thing. plus I think last week was a straightforward week. the first couple weeks of june were pretty good too and then things took a turn for the worse so I need more time to test this out.

  3. Awesome post. Your findings on the first 30 minutes are interesting. This is encouraging me to verify my beliefs. Thanks.

  4. I need to spend some time verifying the study, linnsoft’s homework page has very different results. I think I have something wrong but I can’t find it.

    http://www.linnsoft.com/homework/AvgBracketRangeVol3.JPG

  5. In his trading room this morning, FT71 articulated well why he scales out. Partially paraphrased, he said, “I scale out to pay to reduce risk. It’s as if I’m buying insurance. I reduce my theoretical average, which reduces risk. I’ll always give up some profit in order to reduce risk. That’s how to stay in the game.”

  6. Jeff,

    Thanks for sharing that. I think one has to be careful trying to use FT’s method because it works for him because he has a large account and has very small risk. I’m willing to bet his first scales have a R:R at least 1:1. so he can do that.

    If a 2 lot trader tries that using a 2-3 pt stop, it might not work. It would be great if FT could elaborate on that and how it relates to small retail traders doing 2 lots.

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

(required)

(required)