In the previous article I outlined why top and bottom picking is a topic we inevitably have to deal with as traders and described why it gets such a bad rap. In this posting I will cover off the anatomy of a top/bottom and how to trade it.

The Anatomy of a Top / Bottom

A top of bottom is basically the end of a secondary swing. It might also signal the end of a primary trend, but we have no real way of knowing if the top or bottom is the end of the trend until the market tells us (actually this is not quite true – we do know where the real bottom of the market it is – it is where the price goes to zero). So when I think about picking tops and bottoms, I am usually thinking about the top or bottom of a secondary swing and not trying to guess the top or bottom of a primary trend.

Given that the top / bottom of the primary trend is unknowable, all we can do is just need to keep plugging away at trades on the assumption that the primary trend will continue until the trend bends at the end. If the sentiment / fundamentals / technicals are all starting to look a bit sour, then we should lighten our risk exposure, perhaps switch to swing trading and rely more on tools like trailing stops and taking regular profits. None the less, until you see the primary trend actually bend big time, you don’t know where it will end.

Tops and Bottoms of Secondary Swings

Secondary swings capture the ebb and flow of buying and selling in the market and represent the short term expectations and uncertainty about the primary trend. Compared to picking the tops and bottoms of primary trends, picking the tops and bottoms of secondary swings is more “doable”.

A top or bottom of a secondary swing is actually quite an emotional event which tends to stick out like a sore thumb on the chart. You know it when you see it as it consists of three phases:

• The Lead Up;
• The Climax and Collapse; and
• The Recovery

The Lead Up

Normally, the market undertakes a random walk around the primary trend. The price just “wobbles” around the trend. The thing about about random walks is that long periods where the price continues to make new highs or new lows are very unlikely.

In statistical terms a “run” is where we see a set of similar events repeated one after another. For example, if we are flipping coins it might be a long series of consecutive heads in a row. The statistical chance of that run being a long run diminishes quite quickly. For example, if we flip a coin a 100 times, on average there will be 25 times where heads appears twice or more times in a row, 12 – 13 times where heads appears three or more times in a row, 6 times where heads appears four or more times in a row, 3 times where heads appears five or more times in a row, etc. Just like flipping coins, random walks also rarely display long runs of higher highs or lower lows. Now real markets are not a coin flip. Sometimes they do have a similar behavior and it is statistically unlikely to see runs of sequentially higher or lower bars. But unlike random walks – they do tend to exhibit periods of autocorrelation and have abnormally long runs where they are driven by irrational people and not coin flips.

So here is the trick – if you see a run of 6 or more bars in a row – the odds that the market is random walking is extremely low. The market is most likely to be in the lead up phase and is building momentum towards the top / bottom of a secondary swing. If you like to use oscillators, you will recall that I said that RSI is a useless indicator because the standard reading above 70 or below 30, should indicate over bought or over sold. What you need to realize is that a reading above 70 or below 30 for less than 6 bars indicates that the market is temporarily over bought / sold and the price will revert to the mean. If RSI has been above 70 or below 30 for more than 5 bars you are in the lead up phase for the top / bottom of the secondary swing.

At this stage you really should be thinking about getting on this run away train and going long or short. If you see a run much longer the 6 bars – say 9 or so bars – for god sake do not step in the way of the steam train or you will be flattened. You should be waiting for the climax and getting ready to trade the reversal.

It is difficult to know what the actual length of the lead up will be before the price climaxes and collapses. The usual signs to look for are either:

• The price action starts to approach a major high or low (e.g. a 20 day high / low, a major support resistance line or a historical high / low, etc); or
• The run of higher highs or lower lows has been going longer for 9 to 13 bars.

This second point is one I picked up from Tom Demark’s books on technical analysis. His rule of thumb is based on observations on numerous charts and lead to the creation of his TD Sequential indicator (which incidentally back tests very nicely).

The Climax and Collapse

Towards the end of the lead up phase, the liquidity in the market starts drying up and prices start to move parabolically. This is because the buyers are starting to run out of sellers (in the case of a top) or sellers are starting to run out of buyers (in the case of a bottom). At this stage the price is really starting to spike. Finally, the last buyer buys (near the top) or the last seller sells (near the bottom) and the price collapses. When this happens what we see on the chart is one of two formations:

• A single candle with a very long wick (known as a “hammer” or a “shooting star” in Japanese candle stick jargon or “pinbar formation” in the james16 forums);
• Two candle sticks of roughly equal length which are almost parallel to one another and are in opposite directions (known as a “dark cloud cover”, “piercing pattern”, “tweezer top” or a “tweezer bottom” in Japanese candle stick jargon). These formations are actually no different from the single candle formation described above spread over two candles. If you take the open of the first candle and the close of the second candle and the highs and lows, we get a hammer or shooting star formation.

As soon as you see a hammer, shooting star or tweezer form, you should be taking profits as once the top has been hit, the price collapses quite quickly. At this stage the number of players in the market willing to buy after a top collapsed or sell after a bottom has recovered is thin. As a result the liquidity is still low and prices continue to spike in the opposite direction. Just how far the price will collapse is hard to say. A rough rule of thumb is the price will collapse no more than 50% of the length of the run that led to the top. If you are quick on the trigger you can trade the reversal to the 50% point. However, it is a quick hit and run sort of thing though, so you need to take profits quickly and not be greedy on these trades. Again – forget about trading the reversal if the price has moved more than 5 bars down – it is too late – you missed your chance.

The Recovery

At some stage, traders will start coming back into the market because they see value in buying after the collapse or wish to resume selling after the collapse. This sometimes results in another trend forming again, with the market heading for a double top /double bottom and collapsing again, this time much further, or the trend resumes quite strongly and the market makes new historic highs a lows.

Parting Thoughts

Writing this article on tops and bottoms has been quite helpful for me. Even though I have been trading the run up to hammers, shooting stars and tweezers for years, I had never really understood the market dynamic well until I thought about the role of liquidity in forming the top and bottom, the statistical unlikelyhood of long runs as a determinant for picking a top/bottom formation process in action and the ease of picking tops/bottoms in secondary swings vs. the unknowability of trading the primary trend. Anyway, till next time.

This entry was posted on Sunday, April 20th, 2008 at 9:59 pm and is filed under Technical Analysis. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.