RELATED TOPICS --> | Lowry's 90% Days | Buying Demand and Selling Pressure | The A/D Ratio |
                               | See latest TroughFinder Alerts at the HeadsUP Blog |
                               | *NEW* How to interpret and use the TroughFinder Chart |
                               | *NEW* How to use the JBAR report for risk-reduced profitable trading |

We have seen on previous research that the beginning of new bull market gains normally constitute a large chunk of overall gains for that bull market. That is because the markets tend to "rebound" from extreme troughs with enormous velocity.

The problem is of course that no-one trusts these rebounds. If the corrections have been steep, bad news still prevails and more than likely a recession is still in play. Sentiment is at an all time low. Under these conditions, who wants to commit funds to the stock market for fear of catching a falling knife? Of course, by the time the markets have risen enough for everyone to feel comfortable committing funds to them, the party is almost over!

But what if there was a way during market corrections, where you could "call the bottom" with a fair degree of confidence? What if such a system, although not perfect, still makes pretty good bottom calls even if it is a bit premature or late? Surely such a system would be of enormous interest to investors?

Well, we have such a system. We know that is a pretty radical statement and yes, conventional stock market wisdom always states that its impossible to pick bottoms and you will "land up catching falling knives", but that sage advise holds true for only the uneducated investing public. The knowledgeable investors, technical analysis firms and quants know otherwise.

We have already given you some background data into the psychology of stock market corrections and commencement of new bull markets and the characteristics of market-breadth data during these periods in our "Up/Down Volume and Lowry's 90% Days" and "Buying Demand and Selling Pressure" research papers. Lowry's research, the oldest Technical Analysis firm on Wall Street, makes it its business to use this type of research to the benefit of large fund managers and investing institutions and have remarkable accuracy and track record in calling bottoms and new bull markets.

At PowerStocks Research, we have been analysing JSE market data going back to 1997 to see if we could devise a simple fully mechanical system for using market breadth data, particularly advances, declines, up-volume and down-volume to pinpoint stock market correction reversal points.

We already had the beginnings of a system we have been using very successfully with our subscribers for 6 months, but with the new volume-based breadth data recently available at our disposal we knew we could do better, and more importantly we now had the data to do a serious 12 year backtest.

After several months of analysis and testing, we found the system we were looking for and in September 2009 introduced our TroughFinder signalling system for picking out market troughs.

TroughFinder generates two categories of reversal signals, namely Major signals (for large corrections) and minor signals (for the many small corrections/rebound combinations that will happen on a bull or bear market). There are obviously far more minor signals (one every 3-6 weeks) than major signals (one every 2-4 months.) The purpose of this document is to introduce you to the results of TroughFinder for major signals and then to conclude with a visual demonstration toward the end of this document as to how our system works for major and for minor troughs.

Note that we think we have come close to perfection on picking bottoms to generate BUY signals but when to SELL is another (as important) story we DO NOT cover here. This is a harder problem to solve as stock market peaks happen gradually whereas troughs, driven by panic psychology and greed when prices become ridiculously low happen much more rapidly and are easier to time. The good news though is we are close to completing PeakFinder which will assist with the sell signals. In the meantime our JSE SwissClock provides for excellent SELL signals to use in combination with TroughFinder. But we will leave that for another research note - let's focus on finding stock market bottoms on the JSE - the first step to a solid investment return.

Whilst the detailed mechanics of TroughFinder are proprietary to PowerStocks Research, some of our savvy, more experienced subscribers  have a pretty good idea how it works after using it for a while. TroughFinder has two condition-signals it constantly monitors, namely the A-Signal and the B-Signal. The A-Signal needs to meet a certain threshold for TroughFinder to "become interested" in a situation. The larger and more intense a certain sell-off in the market, the bigger the A-Signal. When determining the value of the A-Signal we look at Decline/Advance ratios (number of declining shares versus advancing shares), Downside Volume ratios (volume of declining shares versus that of advancing shares) and rolling averages of these values. We also look for the presence of Lowry's 90% and 80% downside days.

Once TroughFinder has an A-Signal, then this tells it that "This is a decent enough sell-off, start looking for increased demand, buyers rushing in to snap up bargains and Lowry's 80% and 90% upside days" that mark a classic reversal. This time we start looking for certain Advance/Decline ratios, Upside Volume ratios and rolling averages of these values. These all combine to form the B-Signal and sooner or later the B-signal hits a threshold and provided this occurs within 10 days of the the A-signal we have our event/s calling the market trough.

The A-signal followed by B-signal concept is the same for major and minor TroughFinder signals, it is only the thresholds we use that change. This means less major signals than minor signals but it also means better accuracy and confidence with the major signals than the minor signals.

The system is simple - a large, sustained and intense enough sell-off accompanied by a large and sustained, intense enough buying spree marks your stock market bottom. The rules and maths behind these signals we have developed is purely unsubjective. Any multiple of people following the mechanical rules will arrive at exactly the same result - each and very time.

We put TroughFinder through the mill with over 12 years of daily JSE data, constituting over 3,000 individual trading days to test its robustness. In the period December 1997 through to August 2009 (just under 12 years or 140 months) TroughFinder generated 74 major signals in 39 distinct periods of correction on the JSE. That amounts to one signal every 2 months (60 days) on average for contractions that occurred once every 3.5 months (100 days) on average.

The success stats (major signals only) are as follows:

37 signals were "on the money" calling the trough less than 5 days after it occurred (50% of signals)
11 signals were "premature" making a call up to 10 days before it occurred (15% of signals)
15 signals were "late" flashing 10 days after the actual bottom (20% of signals)
 5 signals were "booster" signals midway through strong bull markets (7% of signals)
 6 signals we "wrong" and preceded further corrections (8% of signals)

With regards the 11 "premature signals", they were still pretty darn good at calling a bottom even if an investor that played the signal lost up to a further 10% before the real bottom hit. The ensuing bull runs still ensured a healthy dose of profit to completely neutralise these initial draw-downs.

With regards the 15 "late signals", once again they were close enough to the actual trough to still classify as "incredibly fortunate" as they latched onto eventual market rises that would have booked tidy profits. We also see that 5 of these signals occurred half way through existing bull markets that we had latched onto from previous signals, serving as powerful confirmations and additional thrust to a strong run.

You would have encountered a severe loss with the 6 "wrong-way" signals that for the most part were "bear-traps" (bear market rallies).

So think about this : 50% of the time you would have encountered incredible timing, and 35% of the time you would have encountered "pretty darn close" profitable timing and 7% of the signals were confirmations of a large bull run that would have encouraged you to put more in the stock market. All in all 92% of the signals would have yielded a healthy profit!

TroughFinder did not catch EVERY major trough and/or bull market beginning but it latched onto at least 75% of them. And when it latched you were assured with a high-confidence event.

The chart below shows the last 12 years with the A and B signals associated with major TroughFinder signals. Note that about 15 periods where A-signals (pink areas) were showing were not accompanied by a corresponding B-signal within 10 days, yet some of these periods still marked pretty good troughs (these were NOT included in our accuracy analysis in the previous section as we do not view them as valid BUY signals). The chart shows the JSE to a logarithmic scale for better viewing of the extent of the sell-off's (A-signals) and subsequent recoveries. Click on the chart for a much larger 100KB version to open in your browser to inspect individual signal accuracy.

You will notice that many signals were "bunched together" as part of a particular consolidation in the JSE and as such the chart above represents 39 distinct periods where "opportunity" presented itself to us and not 74 (the number of signals). If we worked on the basis that the first signal is our best signal (the signal we act upon to enter a trade) then the chart will appear as below (once again click on it to get a bigger 60Kb image to inspect the signal accuracy)

The success rates are equally impressive:

23 signals were "on the money" calling the trough less than 5 days after it occurred (59% of signals)
 6 signals were "premature" making a call up to 10 days before it occurred (15% of signals)
 3 signals were "late" flashing 10 days after the actual bottom (8% of signals)
 4 signals were "booster" signals midway through strong bull markets (10% of signals)
 4 signals we "wrong" and preceded further corrections (10% of signals)

A lot of these stats mirror those of the complete set of 74 signals, except that the "on the money" calls represented 9% more success rate and "late" signals were much lower (due to these being the first signals we saw and acted upon). Overall success rate (in terms of profitable trades) was 90%. Three of the 4 wrong signals were when we got caught in bear-traps in the great 2008 crash.

Every single successful signal (90% of them), even those that were premature and suffered a small draw-down before ground zero was reached and even those "boosters" midway through bull markets, got you in "on a significant dip". If you set your initial stop-losses appropriately upon entering these trades you could neutralise the effect of the 10% of trades that may go against you.

So far we have shown you how TroughFinder finds major market bottoms. But the signalling system we have devised for TroughFinder can also detect intermediate (smaller) troughs. The A-Signal and the B-Signal are compiled by us daily and alerts issued for major and minor troughs to the PowerStocks HeadsUP! alert blog for our subscribers. There is a signal about roughly every 2-3 weeks. In order for you to understand how to interpret the signalling system, let us go into more detail on the A and B signals.

The A-Signal represents cumulative selling intensity as measured by daily Decline/Advance ratios and Downside Volume ratios. Points are allocated to this signal line when days are encountered when either of these ratios exceed certain thresholds. Large ratios representing Lowry's 90% and 80% days attract additional points. After 12 trading days, the points expire as the opportunity presented by the selling wanes, and gets eroded by intermediate buying. You can interpret the A-Signal as your "measure of opportunity building up for a reversal". The higher the A-Signal the bigger the opportunity (and larger the sell-off)

The B-Signal represents cumulative buying intensity as measured by daily Advance/Decline ratios and Upside Volume ratios. Points are allocated to this signal line when days are encountered when either of these ratios exceed certain thresholds. Large ratios representing Lowry's 90% and 80% days attract additional points. The points allocated to the B-Signal line expire the following day, unless the following day also generates B-Signal points in which case the points are added together (to reflect the power of back-to-back or "double-days".) In addition, the points allocated to the B-Signal line are deducted from the A-Signal line the following day, as the buying surge would have diminished the "opportunity". You can interpret the B-Signal as a market reversal attempt if a A-Signal is present, or as a powerful surge in bull market momentum if the A-Signal is not present. During a market decline, when the A-Signal is showing, the size of the A+B-Signals represents the size of the potential market reversal.

A and B signals representing selling and buying intensity for the last 12 years are shown below in the log-scaled chart of the JSE. You can see that all "major" stock market corrections are marked by A-signals of intensity 10 and more.

A large A-Signal followed by a large B-Signal represents intense selling pressure suddenly accompanied by intense buying demand - the classic reversal signal as shown in our "Up/Down Volume and 90% Lowrys days" research on Bear Market bottoms and new bull markets.

We are busy cataloguing each and every one of the 39 "first come first serve" signals discussed at the end of section 5 for our upcoming book and dissertation on this trading system, but let us examine three of them, namely signals # 4,5 and 6 in the Feb-May 2000 correction

1.  Intense selling starts, no buying of consequence in sight
2.  Intense selling frequency increases, evidenced by more A signals. B signal of 2 marks trough
3.  Panic selling, high intensity of 6 followed by "calm before buying storm" lull
4.  Strong B signal, strength 3 marks major trough
5.  Powerful wave of B signals thrust market higher
6.  First A signal for a long time after powerful series of B signals marks exit
7.  Heavy selling of intensity 3 drives market down a slope
8.  A-signal of 4 marks panic trough but small B signal of 1 keeps us out this bear trap.
9.  Crash-class panic selling intensity 8,9 marks major "waterfall" bottom
10. Back-to-back B signals mark trough 3 days after ground zero
11. Another post-crash selling bout followed by B signal marks trough

So you can see there is not much to it. After following these charts and making a few trades it will become like 2nd instinct. Note B signals of strength 1 are fine for trading post-crash but during a crash we like to trade only on B-signal strengths of 2 and above (or two back to back signals.)

The signalling system, as we publish it on WJP for our subscribers is shown below since the last major trough in February 2009.

Daily signals are represented by vertical bars. The A and B signals are stacked on top of each other and thus the height of the vertical bars represent the strength (intensity) of the overall A+B signal. Your BUY signal is depicted on any day when you have both A and B signals showing.

You can see that selling intensity (red A-Signal) built up to a crescendo in the last major 03 March trough to the left of the chart. On 26th February buyers rushed in, triggering thresholds that resulted in the B-Signal tripping. You can see the first vertical green bar peeking out on the top of the red bars on the left of the chart. This was our first sign we were nearing the bottom and was quite a strong signal, measuring an intensity of 10. As you can see the signal was a couple of days premature but even if you acted on it you would have an amazing story to tell your colleagues!

To improve accuracy and avoid false signals, we like to classify major signals (depicted by the orange arrows) as meeting all the following criteria :  

1. An overall A+B signal strength of 6 or more
2. A minimum A signal strength of 4 (adequate amount of sell-off and opportunity)
3. A minimum B signal strength of 2 (adequate intensity of the buying demand marking the reversal)

These "major signals" are the ones used in our 12-year backtests covered at the beginning of this section and are depicted by bright orange arrows for the test period in the above chart. As you can see these were all EXCEPTIONAL signals.

As you can see from the chart, there were many excellent smaller signals that did not meet the major signal criteria. After the market correction had settled and the JSE started trending up there were brilliant minor signals on 4th May (right after the 2nd major orange signal), 15th and 19th May and 17th,19th August.

Of all 21 signals depicted above only the one on 1st June was wrong, but note that the A-signal was only 1 point and the B-signal 2 points for a total signal of 3 points. We advise against acting on signals where the A-signal is less than 2. Using this rule, NOT A SINGLE SIGNAL ABOVE WOULD HAVE RESULTED IN YOU LOSING MONEY. At worst you had to wait for the market to base itself for 1 or 2 weeks but 2 weeks after every signal resulted in a major market move upwards.

Note that as a bull market gets under-way and volatility drops, we will be willing to accept smaller and smaller signals. As you can see from the chart in the last 3 months, B signals with a strength of only 1 are generating very nice signals.

One final note - sometimes an A-Signal is not accompanied by a B-Signal yet it still accurately pegged a market trough. You must always view the A-signal as "opportunity" and many of our subscribers trade on A-Signals of 3 or 4 more to "scalp" the market. They look at the Advance/Decline and Up/Down volume charts on WJP (sample shown below for advance/decline data) to find their own signs of returned buyers (ie accepting lower thresholds of returning buying demand than TroughFinder would) Some of the minor signals we spoke of above are shown below with arrows.

The signal with the big green arrow marked "4" was a TroughFinder major signal that corresponds to Signal-22 in the section right at the top of the page (the 11 year backtest). You can see the A-Signal was generated ON THE DAY of the trough (green arrow on chart on right) and the confirmation B-Signal 5 days later (chart on left). Also shown on the above charts are 4 other smaller signals that ALL corresponded with localised troughs.

You can view the A+B combined signal strength, represented by the height of the vertical bars in the chart as the "thrust" imparted to the market - the bigger the push, the more the market moves up. New bull markets require sufficient initial thrust to allow them to "escape earth's gravity" Martin Zweig mentions this constantly in his book (with particular reference to the A/D ratio), but it is only now with the TroughFinder chart that you can visually appreciate this investing greats' wisdom.

Even when not in a market correction (ie no A-Signal showing) the presence of B signals act as confirmatory "booster thrusts" to a strong bull market.

TroughFinder-II, launched on 15th September 2009, introduced a C-Signal, which is also an excellent predictor for minor troughs. C-signals are based on large negative to positive reversals of Net Advance Volume (NAV). NAV is daily upside-volume subtract daily downside volume. If the advancing stocks for a day amount to 200 million shares and the declining stocks amount to 300 million shares changing hands then the NAV is -100 million shares. C-signals are generated when we experience a 5-day period in which we have a NAV of at least -135 million shares reversing to a NAV of greater than zero. These represent large swings from selling pressure to buying demand that are not always detected by a B signal for smaller troughs. C-signals sometimes show when there are no A or B signals and often overlap with  larger TroughFinder B signal alerts. The chart below shows how the A-signal (red), B-Signal (green) and C-signal (yellow) are represented on our TroughFinder Signal Chart available to our paid subscribers.

You can see that C-signals in almost all instances, flag a local trough within a remarkable 1-2 days. On 31 July, we had no A signal showing but some B signals popped up. When we saw a B signal pop up on the same day a C signal popped up we knew we had a high confidence event. As you can see this was an excellent signal. On 04 July we had a large-ish A signal of 4 showing and on 05 July a C signal popped up - one day after a major trough. This was a stunning bullish signal.

TroughFinder alerts and updates are posted daily at the HeadsUP blog (see below link)

RELATED TOPICS --> | Lowry's 90% Days | Buying Demand and Selling Pressure | The A/D Ratio |
                               | See latest TroughFinder Alerts at the HeadsUP Blog |
                               | *NEW* How to interpret and use the TroughFinder Chart |
Make a Free Website with Yola.