Making of a method

[game]

Having covered four months of the first 90 mins of the day, the statistics show the hypothesis to hold promise.

My task now is to mold the hypothesis into a rigorous set of rules so that there is very little discretionary decision making involved when it comes to executing the trades in real time. The creativity is needed during the backtesting process.

I am keeping my first strategy specific, covering only one variant of the larger hypothesis. Once I have shown the ability to execute it correctly, I will use the confidence gained from the proven hypothesis to construct and test new variants. Hopefully, there will come a day when there are enough tested strategies to participate in a wider space with confidence and freedom.

For anyone who is reading this and considering developing a trading plan, I am following the process detailed here:

http://www.traderslaboratory.com/forums/wyckoff-forum/15535-developing-plan-trading-journal.html

 

[dbphoenix]

Quote from game:

Having covered four months of the first 90 mins of the day, the statistics show the hypothesis to hold promise.

My task now is to mold the hypothesis into a rigorous set of rules so that there is very little discretionary decision making involved when it comes to executing the trades in real time. The creativity is needed during the backtesting process.

I am keeping my first strategy specific, covering only one variant of the larger hypothesis. Once I have shown the ability to execute it correctly, I will use the confidence gained from the proven hypothesis to construct and test new variants. Hopefully, there will come a day when there are enough tested strategies to participate in a wider space with confidence and freedom.

For anyone who is reading this and considering developing a trading plan, I am following the process detailed here:

http://www.traderslaboratory.com/forums/wyckoff-forum/15535-developing-plan-trading-journal.html

Just a reminder that there are only three strategies: breakouts, retracements, and reversals. Keeping that in mind may help you avoid wandering off into the weeds as you start honing and consolidating.

As for "rigorous set of rules" and "very little discretionary decision-making", don't be too hard on yourself if you find that this doesn't work as well as you thought/hoped it would. If you have a "rigorous" set of guidelines, you will have the freedom to be as discretionary as necessary. Granted that fear may still be an issue and may continue to be an issue, being too rigorous and even mechanical may provide so few trades that the whole effort will seem to have been a waste of time.

If you feel that you're becoming "tuned in" to traders' behavior and the upward and downward pulses that result from their transactions, try loosening up a little and flow with it. After all, it's only money

 

[game]

Quote from dbphoenix:


As for "rigorous set of rules" and "very little discretionary decision-making", don't be too hard on yourself if you find that this doesn't work as well as you thought/hoped it would. If you have a "rigorous" set of guidelines, you will have the freedom to be as discretionary as necessary. Granted that fear may still be an issue and may continue to be an issue, being too rigorous and even mechanical may provide so few trades that the whole effort will seem to have been a waste of time.

After completing 4 months worth of backtesting, I ran a rough forward test for the month of March. I say rough because the speed was set to 10x and there was no prep work done for the session (neither was there any prep done for the backtest - just looked at patterns). The results were sobering.

While there were some execution errors in doing the FT, the plan largely, was followed. Though my confidence has dampened a bit, I am aware of the backtesting clearly showing that a Reversal trade after a re-test of some sort is a good idea. On the flip side is the disaster performance from the March FT (much worse than BE). The Daily chart shows that the movement during March was basically flat with very little daily range - so that, probably may have a lot to do with the performance.

Scanning charts for patterns during backtesting can be a cherry picking paradise. The results are close to optimum. I forward tested the best days from the backtest and managed to execute the right trades. That tells me that the backtest results were not entirely fictional.

It was interesting to read on ET about the 123 and 2B setup. Those are the basics of my plan, although I arrived at the hypothesis by reflecting on the best balance of price and information risk.

But the issue is that situations resulting in a clear substantial swing, followed by a clear substantial Reversal effort, followed by a clear substantial continuation effort and a rejection, are not very common. I would like to have a more complete Reversal strategy to counter both poor market conditions and the generally low frequency of good re-test set ups.

Perhaps with some more observation I can come up with Reversal situations containing compelling opportunities that occur with more frequency than the basic re-test scenario....

 

[dbphoenix]

While I'm not optimistic about your efforts to "come up with Reversal situations containing compelling opportunities that occur with more frequency than the basic re-test scenario", I don't want to discourage you from trying. If you're determined and you don't give up easily, you may just come out the other side. Otherwise, you may become defeated for no particularly good reason other than the choice to wander down the wrong path.

I suggest that using the so-called "123" and "2B" as the bases for your plan is an error since they are little more than features of the landscape and have no particular meaning in and of themselves. Unless reversals take place against reasonably important support and resistance, they are far more likely than not to draw the trader into an almost continual state of trading counter-trend. Wyckoff figured this out more than a century ago. Dunnigan, Ross, Sperandeo and others merely re-affirmed it, though they seem never to have figured out that their problem had to do with ignoring support and resistance.

There are other ways to find reversal opportunities. One is to change your bar interval so that you're looking at different levels of support and resistance. Making it smaller will of course mean that the swings you're trading will be smaller. This may suit you. Making it larger will mean that the swing will also be larger. But the waiting may make you impatient. Another is to learn how to trade ranges since they occur more often than trends do. However, trading ranges means having to give up the safety net of the retest since you have to take the trade as soon as the limit of the range is reached (ranges are rarely accommodating enough to provide retests). If you don't want to do that, then you may have to begin monitoring a variety of instruments and trade those that are on the springboard to providing a trading opportunity during the coming trading session.

 

[game]

Quote from dbphoenix:

I suggest that using the so-called "123" and "2B" as the bases for your plan is an error since they are little more than features of the landscape and have no particular meaning in and of themselves. Unless reversals take place against reasonably important support and resistance, they are far more likely than not to draw the trader into an almost continual state of trading counter-trend. Wyckoff figured this out more than a century ago. Dunnigan, Ross, Sperandeo and others merely re-affirmed it, though they seem never to have figured out that their problem had to do with ignoring support and resistance.

I am unclear about backtesting while incorporating context elements of important S/R and long term trend. If the right path lies in the direction of observing trading behavior around important S/R levels, how do I backtest for this?

I arrived at 123 and 2B by first observing in sim mode and then forming a hypothesis reasoned with info/price risk and the story about trading behavior around these patterns. The backtesting has shown that trades around these zones either work or not. There is little hesitation and the heat on winning trades is generally quite low.

I do not want to go down the wrong path for the sake of developing persistence. So I am taking your advice to reconsider this approach and trying to separate what to discard and what to keep from this backtesting experience.

Quote from dbphoenix:

There are other ways to find reversal opportunities. One is to change your bar interval so that you're looking at different levels of support and resistance. Making it smaller will of course mean that the swings you're trading will be smaller. This may suit you. Making it larger will mean that the swing will also be larger. But the waiting may make you impatient. Another is to learn how to trade ranges since they occur more often than trends do. However, trading ranges means having to give up the safety net of the retest since you have to take the trade as soon as the limit of the range is reached (ranges are rarely accommodating enough to provide retests). If you don't want to do that, then you may have to begin monitoring a variety of instruments and trade those that are on the springboard to providing a trading opportunity during the coming trading session.

I really like this entire approach of trading intra day swings. It suits my temperament. I like the thought of taking a position during confusion and riding the ensuing move during the first 90 mins. I like focusing on 1 instrument to understand the principles of PA. And I like the potential within index futures to scale up a winning method.

So I am going to keep the instrument and time frame the same. These are my reasons to continue as is. On the flip side, perhaps I am just being resistant to change.

 

[dbphoenix]

Quote from game:

I am unclear about backtesting while incorporating context elements of important S/R and long term trend. If the right path lies in the direction of observing trading behavior around important S/R levels, how do I backtest for this?

Pick any chart from the past, at random. Zoom out so far that you can't see anything useful or informative about any particular day, or, preferably, any particular week. When you open that particular chart, you'll have a starting point at the left edge and you'll have an ending point at the right edge. If it's a futures chart, these points may not coincide with the NY open and close. That doesn't matter. All you want is a starting point.

Now look at what's in front of you. Is price rising or falling? Where? How and where did it start? How far did it go? How much trouble did it have doing whatever it did? Are there any levels where it was turned repeatedly turned back, either to the upside or the downside? Might these be reliable S/R levels? And so on.

Now scroll forward, bar by bar, or let the replay function do it for you. How is price reacting to the context you're hypothesized? Is it respecting the S/R levels you hypothesized? Or is it ignoring them entirely and possibly creating new ones? Is it trending or ranging? If trending, what is the stride? If ranging, what are the limits? And so on.

Wash, rinse, repeat. Wash, rinse, repeat. Whether you replay ten days or twenty or thirty is up to you. You can replay an entire year in this way.

I arrived at 123 and 2B by first observing in sim mode and then forming a hypothesis reasoned with info/price risk and the story about trading behavior around these patterns. The backtesting has shown that trades around these zones either work or not. There is little hesitation and the heat on winning trades is generally quite low.

These "patterns", however, are consistent with any swing high or low, and unless the swing point is against support or resistance, the "pattern" may result in nothing more than sideways congestion, or it may actually be a retracement within a continuation which the trader will miss because he thought it was a reversal.

Naturally these patterns will "either work or not". But the more you focus on support and resistance and the balance between supply and demand when these "reversals" occur, the higher the probability that you will anticipate the direction of the resulting move and profit thereby. If support and resistance are not part of the equation, you're going to be doing an awful lot of stop-and-reverse.

 

[game]

Quote from dbphoenix:

Pick any chart from the past, at random. Zoom out so far that you can't see anything useful or informative about any particular day, or, preferably, any particular week. When you open that particular chart, you'll have a starting point at the left edge and you'll have an ending point at the right edge. If it's a futures chart, these points may not coincide with the NY open and close. That doesn't matter. All you want is a starting point.

Now look at what's in front of you. Is price rising or falling? Where? How and where did it start? How far did it go? How much trouble did it have doing whatever it did? Are there any levels where it was turned repeatedly turned back, either to the upside or the downside? Might these be reliable S/R levels? And so on.

Now scroll forward, bar by bar, or let the replay function do it for you. How is price reacting to the context you're hypothesized? Is it respecting the S/R levels you hypothesized? Or is it ignoring them entirely and possibly creating new ones? Is it trending or ranging? If trending, what is the stride? If ranging, what are the limits? And so on.

Wash, rinse, repeat. Wash, rinse, repeat. Whether you replay ten days or twenty or thirty is up to you. You can replay an entire year in this way.

Thanks for the details. The process above is on a different qualitative level from the backtesting that I was doing, which was basically just scanning for patterns within the 90 min from the open and recording the stats around these patterns. I will go through one day and post here to get your feedback before continuing ahead.

 

[dbphoenix]

Quote from game:

Thanks for the details. The process above is on a different qualitative level from the backtesting that I was doing, which was basically just scanning for patterns within the 90 min from the open and recording the stats around these patterns. I will go through one day and post here to get your feedback before continuing ahead.

I strongly suggest you read Wyckoff's one-year-long analysis of the 1930-31 market (the first pdf listed). Many of those who are attempting to learn a "trading by price" approach read this section once a week. No lines, no patterns, no indicators, no algorithms, no (fill in the blank).

http://www.traderslaboratory.com/forums/wyckoff-forum/3876-basics.html#post36469

 

[game]

This first exercise was more observation and review rather than backtesting. Other than the first Reversal hypothesis based on the retest, at the moment I don't have anything else that can be measured via a backtest for performance. I don't want to create a hypothesis just for the sake of measurement.

So I am starting off with a broad encompassing hypothesis:

The confirmation or failure of past S/R levels carries information while at the same time offering low price risk entries. Thus, these zones offer the most compelling trades via the best balance of price and information risk.

If S/R is key, then becoming familiar with it's anticipation and interpretation should hold the maximum weight. Once I feel familiar with this, I can form a more testable hypothesis.

May 28th:

Daily: The uptrend that started on April 18th has retraced over the last 3 days. However, price bounced off 2962 on the previous 2 days, indicating some sort of support there. If price crosses this level in the upcoming session, it will be a LL, thus enabling the construction of a trendline and channel. Despite the retracement, the long term trend is up.

Price has been climbing since the close of the 24th with a buying wave starting ON and running 15 points before consolidating. Just before the open - Buying has resumed and broken past the consolidation.

S: 3023 (SL from PM)
S: 3021 (Top of the range from PM)
S: 3018 (Bottom of range from PM)
S: 2985 (S from 23rd and 27th)

http://www.sierrachart.com/image.php?l=1375339268657.png

 

[dbphoenix]

This is very good, particularly your noticing how much more effort it takes to reach 36 the third time.