Trading What I See

I’m sure you’ve noticed that I sometimes “cheat” when talking about patterns and trendlines. Most of the rules you find in Technical Analysis books are designed to keep you out of trouble. Rules like not drawing trendlines through price data. And if you follow these rules you will have fairly successful results.

But remember that these “rules” are designed to simplify explaining the market to readers. What I’ve found is that even when the rules are not completely met, the ending result may be the same. A few days ago I pointed this out with a rectangle. Today I’m going to do it with a trend channel. I’ll call this a “What If” channel.

What if the Chinese market hadn’t made a sudden drop before the market opened yesterday? As you can see on a 15 minute chart, the market had already created an upchannel and then gapped outside the established trend. What if the market pulling back inside the channel meant it was going to ignore this event? What if you extended the trendlines and watched them today? What if you caught the top this morning?

I spend most days drawing trendlines, checking Fibonacci measurements, and asking What If? Sometimes the market gives me interesting answers.

After the gap and run to the top of the 15 minute channel, the market moved sideways to down until it hit the bottom trendline. On the 3 minute chart the pattern was an A-B-C Measured Move, with the bottom also occurring at a 127% Fibonacci external retracement. Add the yellow parallel trendline and the fact it was also at the level of yesterday’s close and there were lots of reasons to anticipate a turn. But probably not enough range to encourage an entry.

But looking back at the 15 minute chart, What If today was a bull flag? Or on a 45 minute chart What If we just created a Test of Top from May 23rd? Two questions that point in different directions. Guess we’ll just have to wait for tomorrow and see what setups develop.

Edwards and Magee: Technical Analysis of Stock Trends

The first market report I heard this morning talked about the drop in the Chinese markets, and suggested we might be in for a repeat of the February drop. The gap below yesterday’s low might have hinted at that outcome, but in less than 30 minutes we were back inside yesterday’s range. Oops!

A gap outside of yesterday’s range followed by a reversal sets up the Larry Williams Oops! pattern. As I said in a commentary last year, I don’t trade the Oops! pattern by itself, but will often use the setup as a trading bias for the rest of the day. I want a very strong signal before I will trade against it. Larry says in Long-Term Secrets to Short-Term Trading that

It is the most reliable of all short-term patterns I have researched and traded.

Today was an example of the setup’s power, and the reason for passing any possible short setups today.

After the Oops! pattern triggered (entering yesterday’s range) price completed a 78% Fibonacci pullback of the morning drop. Then it made a 50% retracement of the rally and couldn’t penetrate yesterday’s low. That bounce from support at the 50% level, combined with an upward bias, would be good for an entry at the first green bar for a rally to yesterday’s high.

The mid-day sideways pattern could have been seen as a Test of Top or perhaps even a Spring setup. But Don’t Bet Against Larry. His statistical research is excellent, and I would need a much clearer setup to consider a short here. The Oops! is a longer term pattern than my normal trading, so I would consider a trade against it as counter-trend.

One more 50% pullback to the support of a previous pivot and the market rallies into the close. One of the reasons the Oops! pattern works is because many traders get caught on the wrong side of the market. As price keeps rising, more and more of these traders finally give up and cover their shorts resulting in a day with no pullback greater than 50%. Thanks, Larry.

Larry Williams: Long-Term Secrets to Short-Term Trading

For some reason my data provider (eSignal) posts two hours of trading on holidays. I ignore this extra data (which plays havoc with my trendlines), and will sometimes start my Fib measurements with the day’s open. That was the case today.

Over the weekend Steve asked some questions in comments about initial protective stops, so I’ll use today’s setups to illustrate my answers. First, I always set an immediate hard stop. Where I place the stop will be determined by previous market action.

I prefer using pivots, but will sometimes use the cross of a moving average. Once in a while I’ll use the reversal of the entry bar as an exit. If I can’t find a logical stop that I like, I’ll pass the trade. If the stop would be more than two points from my entry, I will probably pass the trade.

Because of the extra holiday data, I wouldn’t even be looking for a trade near the open. The first clear setup would have been the short entry at today’s high. Price completed an A-B-C with the “C” leg being 62% the length of “A.” A second indication was that a reversal there would form a divergence with the Stochastic oscillator. And you could see that if the trade triggered (first red bar) the setup would turn into a Spring.

I’ll be using the colored Dunnigan bars to explain the entry and stop level so you might want to follow that link first. As we hit the high of the day at a precise “C=62%” the top bar was colored green.

I like the Dunnigan colors, because in real time I can tell when a signal is given. As soon as price extends below the previous bar, the color will turn to red. That is my entry signal (usually executed by way of a stop.)

In this case that top green bar has a range of 0.70. If my entry is one tick below the top bar and my stop is one tick above it, the risk is 0.90 — well below my maximum of two points. The entry in this case would have been 839.60.

What followed was a nice drop with a sequence of red or magenta bars. That means that no bar made a higher high. Assuming you exited at the first change to green (835.60) that would be a four point profit. (I’m not recommending that as a specific exit technique, but there is certainly no reason to exit earlier.)

The 50% pullback to “B” turned exactly at the white parallel trendline (drawn as soon as the pivot at “A” was complete.) An entry at this point would be handled exactly the same as the first trade. The first down bar turned red at 836.70 after touching the 50% level and the trendline. The stop would go above the pivot at 837.50 for a risk of 0.80.

Because the market acted as though it was creating a double bottom, I probably would have exited this time on the first hint of green for a small gain (1.90.) The first trade had a Reward/Risk ratio of 4 to 1. The second about 2.4 to 1. A question I often receive is what is my minimum R/R ratio.

I hate to break the news, but I have no way of predicting how far a move will go. Sometimes trades fail (all too often.) Sometimes a Spring will move to twice risk. Sometimes it will be a larger reversal and produce ten times risk. Some of my patterns have clear minimum targets, but many do not. The actual reward depends more upon your exit technique, and that is possibly the most difficult part of trading.

When I take a position, I’ve looked carefully at the trading environment — how far away is support, what pattern am I trading, etc. I won’t take a trade unless I think there is reasonable room for profit that is larger than my risk. My risk is the only thing I can calculate in advance.

Much of the reward in a trade comes from your trade management. An excellent setup preceding a good move can either make or lose money depending upon your trading style, regardless of what you perceive as the potential reward. The last trade of the day provides a good example.

The correction from today’s high made a nice A-B-C pattern with “C” turning at 62% of “A.” This setup often occurs at a reversal. “C” was also an exact 127% external retracement of the “A” to “B” move, which also occurs at reversals. The bottom was a 50% retracement on a 15 minute chart. A perfect signal, complete with a Stochastic divergence.

Anyone using my entry techniques on that setup, provided they also leave their stop at the original pivot, would have made over four points on the last move of the day. I would have lost money on that trade. The first pullback was too large and my trailing stop would have created a loss.

But that’s just my trading style. Everyone has to develop their own style — one they are comfortable with. I feel that a key to successful trading is being able to take the trades you like without hesitation. And you can only do that if the “rules” you follow are your own.

I wasn’t thinking about Monday being a holiday until I looked at the first hour of trading. I hope most of you made this into a four day weekend. However it’s surprising how regular the movements often are on a narrow range day.

After a small gap, the first hour had a trading range of under five points, and because of the holiday weekend the day’s entire range was only a little over six. The market made an early pullback and then rallied with a small A-B-C Measured Move that completed a larger A-B-C pattern that began yesterday. That gave a double Fibonacci signal for the high of the day.

From the top, a Measured Move down (100%) created the first leg of a larger A-B-C. It ended with a Stochastic divergence at the same time the market tested a pivot from earlier in the morning.

A rally of 50% set up another reversal, and then it was down again in an A-B-C Measured Move to set up a bottom pivot. This was the third touch of a support level shown by the blue horizontal line, the third Measured Move of the day, and it completed a larger A-B-C that ended at a Fibonacci 62%. Lots of tradable setups but unfortunately no range to work with. Have a nice long weekend.

Whenever I enter a position, after setting a protective stop loss I begin looking for my first exit level. My compromise between getting out too early and holding a position too long is to trade multiple contracts.

Sometimes finding a logical point to take some profits can be difficult. I don’t automatically exit at Fib or support/resistance levels, but use them as areas to watch price closely for reversal indications. And sometimes I’ll use points that others may not consider “correct.”

After a nice Spring setup just after the open (complete with a Stochastic divergence), price re-visited a resistance area created yesterday, making a third reversal at exactly the same level. Although there is only a single bottom pivot, I’ll probably start to treat this area as a potential rectangle.

A “correct” rectangle will have at least two pivots at both top and bottom, but notice what happens when the bottom support breaks. There is an immediate pullback to that support, just as often happens with a “real” rectangle. If you’re willing to trade “almost” patterns, that pullback was a nice entry for a short sale.

For those not familiar with the quote used as today’s title, here is a more complete version:

If it walks like a duck and quacks like a duck, you can be reasonably sure it is a duck.

For my trading the same thing applies to rectangles, triangles, and many other patterns. If it acts like a rectangle, I’ll certainly mark off a rectangle measurement for a target even though it doesn’t meet all the normal requirements.

And that rectangle target is where we got a nice reversal today. We eventually moved somewhat lower, but first targets don’t necessarily mean the end of a move. But they are certainly good places to take partial profits.

One of the most difficult things about discretionary trading is deciding how much leeway to give patterns and levels. If you are just starting out, my suggestion is to require very strict adherence to the “rules”, but after you have analyzed thousands of charts you’ll begin to see more “almost” patterns and discover that they often work.

Yesterday I closed the commentary without having any conviction about today’s market direction, but pointed out that continuing moves usually have a large first retracement (50% or more) followed by a smaller second retracement (normally 38%.) Since we closed with a second pullback of 38% there was certainly the possibility of additional rally.

Evidently the market overheard me, but after moving up again it refused to stop at my expected 162% level. Instead it gave a reversal signal using the Fibonacci 127% external retracement. External retracements are created when a move is retraced by more than 100%, and I feel that one of the most tradable reversals comes at 127%.

Think of this setup as being similar to a Spring. The previous top (the resistance) is slightly exceeded, and then price reverses. I like them even better when they are accompanied by a Stochastic divergence.

That’s what happened at this morning’s top, followed by a drop to the support of yesterday’s high. It was also a 50% retracement of the morning’s move. As soon as that pivot was created it was time to draw the lower yellow trendline and it’s parallel. That pivot also creates the rising blue trendline. The pullback stopped at the 78% Fibonacci level just as the yellow parallel was reached.

The next move down would have been difficult to trade because the moving averages are quite flat. There is also that rising blue trendline giving support. But shortly after it was broken the moving averages turned down.

I like to trade pullbacks to a trending moving average, and that occurred at what I have marked as the magenta “B.” The entry would have been just as price broke beyond magenta “A.”

The bottom around 11:00 (Pacific) had two markers — an A-B-C that ended at a Fibonacci 162% and another 127% Fibonacci external retracement of the morning rise. And at the second bottom there was also a small Stochastic divergence.

Drawing another trendline and its parallel gives you the “B” reversal, followed by a third 127% external retracement that matches the last A-B-C at 62%. Lots of multiple signals for entries and exits today, with three good examples of the importance of the Fib 127%.

Usually when I point out a setup, it’s a winner. If you’ve done any trading at all, you know how much easier it is to find the winning entries and exits after the market closes. And it’s easy to miss seeing the setups that fail. That’s why it is essential to either paper trade or trade small size (my preference) when trying new methods.

Today there was a nice Spring setup that failed. But let’s start at the beginning. There were two Measured Moves, first up and then down. A Measured Move is just an A-B-C pattern where the “C” leg is the same size as the “A” leg.

The pullback I have marked as a yellow “B” had a number of near misses as far as trendlines and channels are concerned, but the precise Measured Move was accompanied by a nice divergence for a pullback entry. When I have a cluster of levels that don’t quite match for a pivot I’ll often use the cross of the shorter moving average (13 ma) as my entry.

Then came the Spring. I see nothing wrong with it as a short setup. We broke above yesterday’s high and immediately reversed, and that top would have created a precise Measured Move A-B-C (in yellow.) There was even a small divergence in the Stochastic oscillator.

What do you do when a setup fails? If you trust the setup (and why else would you trade it?), you’re job is to forget the failure and re-evaluate the market. We still have a potentially large A-B-C, but with the divergence at the close it looks as though it may not hit the 162% Fibonacci extension.

But this pullback has just reached 38%, and a continuing move will normally have a large pullback (at yellow “B”) followed by a small pullback (the close.) I guess we’ll just have to wait for tomorrow and the next setup. After all, that’s what daytraders do.