My Trading Plan

Here's a 'detailed' explanation of my trading plan. Please tell me what you guys think of it.

Foreword: The whole thing is based off of the theory that if something is profitable beyond a reasonable area of slippage and errors (such as computer problems) then it will likely be profitable for at least the near future. It's also based off the theory that the market will likely move at least 2 points in either direction from the open in any day before making any huge moves. (The exception to this is a trend day.)

Background: I back-tested this system starting January 1st 2010 and found that it is profitable by 72 points as of this writing.

How it works:

I look at the open price compared to the close of the previous day and decide whether to trade towards the close or away from it. I make this decision minutes before market opens at 9:30 and take a position at as close to open price as possible (usually slippage means one or two tick difference). I then set a 2 point target and 10 point stop and I don't do ANYTHING until either one is hit, at which point my day is done.

Please leave ANY questions of feedback. I've live traded this for almost all of May now and it has been profitable. I am looking for validation, especially from the people with years of experience because I don't know if I'm on the right track or not.

here is a link that explains the concept of "air". If I had to pick one target to trade for it would be air pockets

http://www.mypivots.com/Board/Topic/5507/1/price-bar-overlap

Great to finally understand your air pockets. Another piece of the puzzle in understanding price action.

Interesting thread ... from having developed over a dozen fully automated system (and backtested them and traded the 3 most promising), I have the following comments :

- in general, you need at least a sample of 300 trades for your backtesting results to have any meaningful value

- once you have win/loss %, avg win $$$, avg loss $$$, run a monte-carlo simulation for 1 year worth of trade (in this case, that would 250 trades), watching for the distribution of the max draw-down.
Expect to encounter at least the 25th percentile of max draw-down. Prepare for it - emotionally & financially - can you handle that max draw-down if it occurs on your 1st set of trades ?

- how/when will you decide that your edge is gone ? (I personally do this by selecting a max draw-down somewhere between 5 & 15th percentile)

- will you "trade" your equity curve ? ie., stop trading at a predefined draw-down level, and resume trading after the system shows a meaningful recovery ?

- if your backtesting results show little difference between random direction and carefully crafted direction, what about trading both directions (in separate accounts) ? may-be, using different position size as result of your pre-open analysis ?

dom. thanks for your suggestions.

when you said you've traded the 3 most promising systems, were you profitable and if so, for how long? also how long have you been profitable for as a trader period?

i ask these questions because it seems most of the successful traders on this board are discretionary and im a very systematic logical type of a guy with the whole science/engineering background.

Originally posted by feng456

dom. thanks for your suggestions.

when you said you've traded the 3 most promising systems, were you profitable and if so, for how long? also how long have you been profitable for as a trader period?

i ask these questions because it seems most of the successful traders on this board are discretionary and im a very systematic logical type of a guy with the whole science/engineering background.

1st system - traded about 1.5 month (~70 trades) - started on an 8 loser streak (actually, 2 losers 1 win 8 losers, net -$1500) recovered and went up about $1k net, then took 10 losers in a row (of which 5 same day on the Dubai debt crisis end of November 2009, would have been even 3 more that day if I had not shut down the system mid-day), at which point I was net down -$600 and decided to put the system on hold. I did monitor its performance for a few weeks, it did not recover during that period, and I decided to forget about it.

For the record, this system was indicator based, and used STPLMT entries. I had not backtested it much (only a couple month), main reason being lack of historical tick-data (I am an Ensign user, and this is the weakest point of Ensign IMO).

2nd system - Also indicator based, this one I backtested on about 5 month data (375 trades). I traded it for 3 weeks, the 1st week was good (up $1k) then after a few days I hit a string of 14 losers (1 win in the middle), 5 wins, 4 losers - at that point I was net down -$200 (after 48 trades), I decided to put it on hold because the string of 14 losers was way outside of the backtested history, and I didn't want to experience on new grounds with my real money on the line.

In both cases, I stopped the system way before experiencing the max drawdown that I had decided upon as a "edge is gone" line in the sand.

3rd system - Is my 5th-wave system, there is some info on it on my second post here http://www.mypivots.com/Board/Topic/5777/1/doms-notes

In the case of this 3rd system, I stopped trading it automatically after 4 month, at BE (break-even), not because of its own lack of performance, but rather because its relatively large stop was worrying me so much that it impacted even my discretionary trading (not to mention I would often override the system's trade management)


Re. profitability, I am still fighting with it ... for discretionary was up in April, BE in May (before April, my discretionary trading was only on sim, and I was very nicely profitable - all my issues at this time are in my head, and can be summarized in a single word : FEAR).


I, too, am coming from an engineering background, and to this date I am still convinced that automated trading is the best way to go (for me), as it is the closest to backtesting results (and it scales better - to multiple markets - than discretionary trading). BUT, I know for a fact that is it immensely difficult to develop an automated day-trading system with a forward positive expectancy ... not too long ago, I was even on the verge of thinking it is not doable. Sure, it is possible to create systems that after optimization show to be profitable in the past, but that doesn't mean squat about the future.

At this point, I have ruled out all (bar-based) "indicator" based systems, because bar-based indicators rely on the number of bars from point A to B, and this is essentially a function of market noise (subject to change anytime), regardless of the timeframe (be it min, ticks, volume, range).

Also, it is impossible (for me anyway) to factor in an automated system the same amount of context awareness as we have in discretionary trading.

I think this comment by dom is one of the most important aspects standing in the way of pure system trading success, especially over time. One can only code so much for "context" as he mentions ... and one can only code so much for capturing "tape reading" etc. etc.

Rule based trading with gray matter bridging the gap with some amount of discretion, in my own experience as well as that with tons of other traders I've known well, has been the approach (the blend) where I've seen success ... especially over time. This is especially true for individual "retail" type traders. I have not personally known ANYONE who has created or found a pure system to trade successfully, espec. over time (with ongoing tweaks and all).

I'm familiar with Toby Crabel's work to some extent ... visited with him after a presentation, have his book. He may be doing well with pure system trading. I've also got another buddy that used to trade with him. At one time, I was under the impression he was trading purely mechanically. He was still researching and coming up with systems and adjustments to current systems. [interesting sidenote: Ed Dobson (former owner of Traders Press) couldn't get Toby to let him publish any more copies of his book ... I believe there was only 1 initial run of 3000 copies]

Back to dom's comment at the end of his last post: great observation, something that most folks don't think about or consider in their quest for pure mechanical trading systems!

Originally posted by dom993

Originally posted by feng456

dom. thanks for your suggestions.

when you said you've traded the 3 most promising systems, were you profitable and if so, for how long? also how long have you been profitable for as a trader period?

i ask these questions because it seems most of the successful traders on this board are discretionary and im a very systematic logical type of a guy with the whole science/engineering background.

Also, it is impossible (for me anyway) to factor in an automated system the same amount of context awareness as we have in discretionary trading.

Hi Feng, just food for thought, I also tried figuring out a system for several months focusing on the opening trying to scalp 2 pts off the open, I backtested pretty good results actually just going opposite the 5 min candle just before the open, ex: 9:30 candle close red. go long. my potential stop was much tighter though, 2.5 - 3 pts, it seemed too random so I never actually traded it for real and I haven't looked at it recently because I trade mostly a tick chart now.

Your plan got me thinking again, especially with how wide a stop you are comfortable with. Have you looked at waiting until the price moves 3,4 or 5 points away from the open, and place a trade in the direction of returning to the open. 10 point stop, and if the price continues to move against you 5 points look to exit break even. I back tested the June contract with pretty intriguing results, especially if you can leave a runner and trail. The market seems to like to retest the opening price. It might have to be adjusted on market volatility because the most difficult period was march and early april I think where the range was too small to get an early entry.

Just food for though FWIW

Originally posted by chrisp

Hi Feng, just food for thought, I also tried figuring out a system for several months focusing on the opening trying to scalp 2 pts off the open, I backtested pretty good results actually just going opposite the 5 min candle just before the open, ex: 9:30 candle close red. go long. my potential stop was much tighter though, 2.5 - 3 pts, it seemed too random so I never actually traded it for real and I haven't looked at it recently because I trade mostly a tick chart now.

Your plan got me thinking again, especially with how wide a stop you are comfortable with. Have you looked at waiting until the price moves 3,4 or 5 points away from the open, and place a trade in the direction of returning to the open. 10 point stop, and if the price continues to move against you 5 points look to exit break even. I back tested the June contract with pretty intriguing results, especially if you can leave a runner and trail. The market seems to like to retest the opening price. It might have to be adjusted on market volatility because the most difficult period was march and early april I think where the range was too small to get an early entry.

Just food for though FWIW

i also looked at whether or not the 5 minute bar prior to open was an indication of direction but i found it was totally random (~50%).

however, what you've suggested with the 3,4,5 pts away from open entry might be a really good idea. i dont do runners and trailing stops though but i will investigate key areas like ratchet levels and +/-4 pters.

This plan has now been terminated as it has proven to not be profitable.

Originally posted by dom993

1st system - traded about 1.5 month (~70 trades) - started on an 8 loser streak (actually, 2 losers 1 win 8 losers, net -$1500) recovered and went up about $1k net, then took 10 losers in a row (of which 5 same day on the Dubai debt crisis end of November 2009, would have been even 3 more that day if I had not shut down the system mid-day), at which point I was net down -$600 and decided to put the system on hold. I did monitor its performance for a few weeks, it did not recover during that period, and I decided to forget about it.

New member, first post.

I have been reading these pages for the last few weeks based on a reference from a friend. Seems to be a great deal of information discussed and much to learn. One thing here did catch my attention with the post above. Maybe others can add their perspective too.

dom993 mentioned that in a 1.5 month period he placed 70-80 trades that resulted in a -$600 standing whereupon he placed that system on hold. Other website blogs have suggested not starting this type of trading without a $5000-$10000 initial account. dom993's 1.5 months would imply that he traded 2-3 times a day over that +/-30 business day period that ended down $600 relative to his starting point. That makes me wonder.

Based on your knowledge, and WITHOUT asking for any personal information, what is the average/typical annualized returns (%) that can be expected with the proper application of the prevalent systems being discussed here using a theoretical initial account somewhere of about $50,000?

Not looking for anyone's personal experience so much as what one can fairly expect to obtain in return for the time, risk and efforts expended in this type of investing.

Thanks.

the specific criteria data ur looking..of course noone else here has done backtesting for it. something u need to spend time doing urself.

gaps are good though at fading. just gotta find something that works and be disciplined enough to follow it.