Russell Leslie was an earlier member of Hotcopper MK 1, who had a strong mathematical background , as an early trader I likened him to Einstein ,
Russell only used Excel spreadsheets to analyse and select his trades from,
I collected some of his posts in my early days of trading so as to study his technique
I came to the conclusion that Russell was a very good trader as he would only enter a trade when it fitted all his criteria in the spreadsheet ,and then religiously adheres to a stop loss ,
This is one of Russell posts and he had links to his own web pages , previous management said that he was advertising via his web site and they fell out ,what a loss to them,
It is possible that some people will not be able to understand the main gist with out the back up from his web site ,which has long gone others may glean something from this post
I have recently emailed Russell and asked him if he would consider reposting on Hotcopper again unfortunately he declined .
It is my belief if we have more quality posters we will attract more of the same calibre or better,
I lay no claim to this post as it is all the work of R.L's ,as we called him
Russell Leslie INTRODUCTION - IF ALL YOU HAVE IS A HAMMER….. My spreadsheet is based on the way I trade and the things I want to know. Since my background is heavy on maths and statistics I tend to approach all problems as problems of numerical analysis (if you only have a hammer every problem looks like a nail). My problem is to make money from shares - my solution is to hit everything with a lot of maths (my “hammer” of choice).
Since not every problem is really a nail, my attempts to use a hammer will not always be a success.
If people are interested I will try to update this spreadsheet on a (semi-) regular basis, i.e. whenever I come up with improvements that I think are worth uploading. It is worth noting that I make changes for my own purposes on an almost daily basis. I don’t intend to waste everybody’s time by uploading every little thing that changes. The link to the spreadsheet is (not avaiable now)
A - General Notes 1. Comments are available on most sheets, if you see any red triangles in a square click on these triangles to see my comments on the field's contents.
2. I have tested the spreadsheet with up to fifteen year’s of price history, but with that much data the formatting macros take several minutes to run and the spreadsheet will be around 5 Mb. Also graphing details are difficult to see properly on the Full Chart - all in all it is better to use this spreadsheet with less than three years of active data. 3. The only lower limit on the prices of shares relates to the four decimal place precision of Excel currency variables. In practice this means that shares worth less than $0.01 will tend to graph very poorly. 4. This spreadsheet is a work in progress and I change the indicators that I use on a regular basis. I only ever use indicators that I understand well enough to either program or mark on a chart for myself. I am not comfortable with "black box" indicators - I need to know why a buy or sell signal is a buy or sell signal. 5. All of the indicators that I use where designed for, and work best in, sideways markets and are meant to help you beat the odds in a market where there is no consistent up or down trend. Shares prices that maintain long term up trends (or down trends) will usually not give useful signals to this type of analysis. 6. All implementations of various indicators are in line with my tastes and my tastes may differ from standard sources (e.g. for MMA and Stochastics). When I know for sure that the standard way is better (use of 12, 26 and 9 for MACD, rather than my preferred 10, 24 and 9) I go for the standard way - otherwise I just follow my taste. If the results of my version of an indicator are different from the standard version and I don’t understand WHY they are different then I leave that indicator out until either I can make the results the same as standard sources or I understand why they are different.
7. It is important that you ONLY EVER use indicators to assist you in making your trading decisions - you should not let them make your trading decisions for you. BIOGRAPHICAL NOTE - Just because I say my background is physics and maths does not mean that I am saying that I am particularly good at it. I am a pretty pedestrian physicist and my attempts to climb the peaks of Academe were always unsuccessful. I console myself for my bitter disappointment by making money instead. B - Notes in connection with on the "Full Chart" 1. High, Low and Close are plotted as lines - I don't use opening values (not because they are not useful, just because they make no sense in a line plot). Excel will let you plot with Up/Down bars or even to produce Japanese candles (with OHLC data). But for stocks on a consistent up or down trend there will either be no Down bars or no Up bars and VB will cause the sheet to crash (this error does not appear to be trappable). I have chosen lines as a work around as I can avoid errors in all trading histories. 2. Moving Averages - all moving averages are calculated as simple moving averages. This is not to everyone's taste. There is reasonable evidence in the literature that the simple MA is at least as good (or at least no worse than) the alternatives. I have tried exponential, geometric and harmonic MAs but in the end it was simply my preference to use simple MAs. The periods for the two MAs can be chosen by the user. Currently I have allowed for the plotting of two MAs, one short - usually given as 12 days and one long - usually given as 26 days. I have provided a macro to allow users to change to exponential moving averages if that is their preference. 3. Limiting MA values - I use limiting MA values (highest and lowest MA values in period of interest) as a surrogate for retracement values. I don't have a theoretical justification as to why this often works but it does seem to for share histories I have considered. 4. Linear Trends - The "long period" trends (where "long" is defined by the user - 26 days by default) are calculated from a linear regression analysis and are a bit of an attempt to predict the future. If we assume that shares prices are going to move forward in time on a random basis with a certain average step size then we can have some idea of the likely direction of future price movements based on today's price. The upper and lower trend bar define (very roughly) the 95% probability path for the stock (detailed notes on this are included below). 5. Bought/Sold indicators are marked using data from the Buy and Sell Sheets 6. “Target Buying Price"/"Target Selling Price" are included as a line (depending on whether you are looking to buy or sell). 7. The CBL pivot point and trend value are plotted as lines. The earlier versions of this spreadsheet used price trends to decide whether to plot the CBL high or CBL low value, this was mathematically sound but sometimes produced unexpected results. The current version just plots which ever pair is closer to today’s closing price of the share. 8. The weighted average buy and sell prices are included as lines. 9. Fibonacci retracement levels are included as lines. I have not taken the steps necessary to ensure that these values are rigorously calculated as I did not consider the task important enough. There was room so I included them. Formally, the retracements should be from recent high to preceding low - I just take it from high to low (possibly even recent low to preceding high, which is known not to work). Take retracement levels with a grain of salt. C - Notes in connection with the "80 WD Chart" 1. HLC lines are plotted (see comments on why above). 2. 3. Short and long period Linear Trends are plotted (see comments on why above). The confidence interval of the short period linear projection is difficult to define. I use it only as a trend change indicator in conjunction with the long period linear trend. When the long and short period trends diverge this may indicate the trend in price movements is changing. 4. The short period MA is plotted 5. “Target Buying Price"/"Target Selling Price" is included as a line (depending on whether you are looking to buy or sell). 6. CBL values are plotted as on the Full Chart. 7. Fibonacci retracements are included (see notes above) D - Notes in connection with the “A-D” chart 1. This is volume indicator that is out of fashion now (replaced in most applications by "On Balance Volume"). I find it useful - an uptrending value means money is going into the stock and a down trending value means money is leaving the stock. 2. The MA (blue line) is just there to keep things in perspective and cut down on "noise". 3. A formatting macro makes the zero scale of the graph the average value during the period of interest - this makes trend somewhat easier to see and the scale somewhat more open. 4. Bought/Sold indicators are marked using data from the Buy and Sell Sheets. E - Notes in connection with the “MMA1” and “MMA2” charts (Note: this is NOT the Guppy MMA chart - there is nothing wrong with the Guppy MMA - mine is just different - a matter of taste - the differences are “significant” but not “important”. The moving averages on MMA1 are simple moving averages and on MMA2 are exponential moving averages) 1. On MMA1 2-90 day MAs are plotted in three groups (short - blue, medium - green, long - red). Narrowing of the bands of colour tend to indicate consolidation and presage a breakout in price. On MMA2 2-233 MAs are plotted in various colours. 2. Bought/Sold indicators are marked using data from the Buy and Sell Sheets. 3. An overall linear trend and a 200 day MA are marked as black lines. F - Notes in connection with the “MACD” chart 1. MACD (Moving Average Convergence/Divergence) is a trend indicator (or trend following indicator). A positive value means the price is trending up, a negative value means the price is trending down. 2. The blue "signal" line helps to decide whether you should buy or sell - read a good indicators book to understand this - for some reason some people find this hard to follow. WARNING - false buy sell signals are extremely common - do not blindly follow this indicator (see some more notes below). 3. The difference between the MACD and the signal line is plotted as a red histogram. This makes it easier to mechanically decide when the MACD is signally “buy” or “sell” (if both the MACD and the MACD histogram are negative it is a buy signal, if they are both positive it is a sell signal). 4. Bought/Sold indicators are marked using data from the Buy and Sell Sheets 5. It is possible to use the “Exponential Moving Averages” and “Simple Moving Averages” macros to swap between the form of Moving Average used in the spreadsheet. The change will have its greatest apparent effect on this chart. G - Notes in connection with the “Stochastics” charts. 1. Stochastics are indicators of market strength/weakness - there are lots of them. My presentation and calculation differ from all major sources for these numbers. I am comfortable with that - if you are not then you can rewrite them for yourself. 2. I make little direct use of Stochastics in my trading. This chart holds all of these indicators more as a means of keeping track of how they are calculated rather than because I have any immediate need for them. Presenting 5 stochastics (three “fast” and two “slow”) highlights both their differences and their similarities. H - Notes in connection with the “RSI-MFI” charts. 1. The Relative Strength Index (RSI) and Money Flow Index (MFI) are measures of market sentiment. Values over 70 (or under 30) indicate that the market has deviated from the underlying price trend and you can expect the price to revert to the trend value over time. I - Notes in connection with the “Volume” charts. 1. A “Volume” chart has been included. This chart has been plotted in a non-standard way. I plot with a base that is a rounded down version of the minimum volume (295,000 will be rounded down to 200,000 etc) and a maximum scale that is a rounded up version of the 99% percentile of volume. This range will cover 99% of all trading days. I have included the limits to prevent scale distortions caused by atypical volumes that sometimes accompany option and warrant expiry. On the “Volume” chart I have included short and long period moving averages - minimum, average and median limit lines. J - Notes on Macros 1. ALL FORMAT (macro) - This macro does just what its name says - it applies my preferred formats to all fields and charts. In addition it performs some calculations and places new values for certain parameters on the scratch sheet (e.g. CBL High, Low and Pivot Point). Shares worth less than $0.01 tend to graph poorly. If the stock code begins with “X” the macro assumes that it is an index and treats it accordingly. 2. BOLLINGER BANDS - will replace the two MAs with Bollinger bands. Certain calculated values will not work when the Bollinger bands have replaced the MAs. It is recommended that you use one of the Moving Average macros to put the sheet back into a normal state once you have finished with the Bollinger Bands 3. COUNT BACK LINE (2 macros “high” and “low”) - Output results to Sheet E for use in Sheet A and charts - the results include the Pivot Point of the calculation. 4. EXPONENTIAL MOVING AVERAGES - This macro has been included to allow the user to replace Simple moving averages with Exponential moving averages. The change is most obvious when it is made while watching the MACD chart. The macro makes use of the same time periods that are used by the Simple Moving Averages - the relevant percentage is calculated by the spreadsheet (the percentage calculation is exact - a 9% factor makes for 21.2 days - the macro calculates 21 days exactly). 5. FORMAT (4 macros “AD”, “Special”, “MACD” and “Volume”) - These macros are used by All Format and can also be called separately. They may produce unexpected results when called alone - but in the majority of cases they can function alone on a data set that has already had All Format run upon it. 6. LOAD CSV - will load COMSEC CSV data into the spreadsheet - provided the CSV file has the same name as the XLS file. This has been included only as an example of a “loader” macro, basically each new data format will require the writing of a new macro. 7. RELATIVE STRENGTH INDEX - This macro calculates the starting values for two parameters necessary to the calculation of the RSI and MFI. It only has to be run as part of All Format. Once these values have been calculated, the user can change the RSI period without having to run the macro. The opening value of the RSI will be more accurate after you run the macro but the final value will not change significantly. 8. SIMPLE MOVING AVERAGES - (Formerly Moving Averages) This macro allows the user to return to Simple Moving Averages after using either Bollinger Bands or Exponential Moving Averages. K - Notes on loading data into the spreadsheet 1. The spreadsheet that I have uploaded, loads all of the data into Sheet B. I have written a number of "loader" macros that I use to condition data from a variety of sources for inclusion in the spreadsheet. I have not distributed them because they are too source specific to be generally useful. 2. There is no problem with loading data directly into Sheet B by hand using "cut and paste" (providing the data is in the order, Share code, Date, High, Low, Close, Volume and sorted in ascending date order). After loading data you should run the "All Format" macro. 3. Currently I am following a total of 38 shares and 2 indices (12 open positions, two closed positions, 24 watched shares and the indices as benchmarks). I download my "Watchlist" from Commonwealth Securities, this contains data on the day’s trading activity. I then use a macro in a separate spreadsheet do my "End of Day" (EOD) for me. It feeds the (very limited) EOD data that I need into each of the 40 spreadsheets and each spreadsheet updates itself. When the EOD is finished I open separate sheets that analyse how my open positions are doing and also keep me informed of trends on positions that I may want to open. 4. I know that some people do EOD downloads for the entire ASX and then analyse using a package such as Metastock, looking for "Buy/Sell" opportunities. It would be difficult to do that sort of analysis with my spreadsheet. The spreadsheet works best with around 2 years of historic data (80 days is really the minimum). It takes around 10-20 seconds to do one share on a fast computer (open the file, add the data, format it, save it and close it - you have to close the file because any more than about 15 open one Mb Excel files will crash most Win95/98 systems). 5. As a programming exercise I intend to write a “massive data” handler in Access, if this project is successful I will upload the result to HC and that will be a much quicker “data trawling” tool. If you tried to use my current spreadsheet to trawl the ASX, the EOD would still be running when the market opens the next day. 6. Each spreadsheet comes with a macro titled "Load CSV". This is included as a sample loader only. If you are analysing share "XXX" it will look for a file called "XXX.CSV" in the default directory. The data will be expected to have a header row, a first column with "XXX" in it, the second column must contain date, the remaining columns must have High, Low, Close and Volume in that order. The macro will ensure that the file is sorted in ascending date order. It will find the lowest date in the CSV file. It will search for that date in the spreadsheet, if it finds it will load all data at that point, if it does not find it will start forward of the next lowest date. 7. The "Load CSV" can be expected to correctly load data from the download section of Commonwealth Securities - anything from another data source would be a bit of a gamble. “Load CSV” is only an example - it is expected that the user will write their own macros to match their data source or just use “Cut and Paste” to enter data. 8. I am currently using the spreadsheet in conjunction with an ACCESS database. All stock data is actually stored separately from the spreadsheet and only loaded when I want to look at a particular stock. This is very easy to do with Access 2000 and somewhat more complicated for ACCESS 97. L - Notes in relation to linear regression trendlines to project prices forward in time. 1. One of the criticisms that are made of technical analysis is that it attempts to predict the future. In a sense this is true, but I believe the criticism is based on a misunderstanding of what we are trying to achieve. 2. If you want to know what is going to happen in the future, a good starting point is what happened in the past (the sun has always risen in the east, I expect that it will rise in the east tomorrow). This is called induction. You can't prove that induction works, but it is often a reasonable place to put your money. 3. My spreadsheet draws three trend lines on the Full Chart - A - The linear forward projection of today's price (based on the closing prices for the last X days, where X = long period entered on Sheet A, the default value 26 days). B - The linear forward projection of the high price (based on the high prices for the last X days displaced upward by twice the average daily volatility). C - The linear forward projection of the low price (based on the low prices for the last X days displaced downward by twice the average daily volatility). 4. These three lines define my “two sigma” confidence interval. For a perfectly liquid stock that was trending strongly the price of the stock would be between these limits 95% of time. The “Confidence” spreadsheet and macro can be used to determine a more realistic value for a real stock in the real world. 5. The same three lines are drawn on the 80 WD Chart and an additional three lines for similar projections based on the short period are also drawn. The confidence interval of the short period linear projection is difficult to define. I use it only as a trend change indicator in conjunction with the long period linear trend. When the long and short period trends diverge this may indicate the trend in price movements is changing. 6. Once again people seem to misunderstand what “confidence interval” means (sometimes they really aggressively misunderstand what this means). If I am following 20 shares, on average, one of these shares will be outside the bounds of my 95% confidence interval on any given day. Sometimes there might be no shares outside the confidence intervals. Sometimes there might be five outside my confidence interval. Just occasionally they will all be outside the expected bounds. In general terms it is not possible to predict such events, but they always remain possible. 7. When I draw the lines and a share falls outside the lines it doesn't mean the technique is wrong - it just means the technique is operating as expected. I have only drawn an approximate confidence interval. Some people seem to think that when I say that "I am 95% confident of something" that I am saying "I know for sure what will happen", instead I am saying "I know I will be wrong around 5% of the time". I am not likely to be wrong less than 5% of the time, hopefully I will not be wrong too much more than 5% of the time. 8. If changes in share prices were really random the distribution of prices projected forward in time would fall within the two outer forward projection lines 95.7% of the time (the statistics of large numbers of random events are very well understood in the physical sciences). Changes in share prices are not random, but they do have a large random element (as a share increases in trading volume and liquidity the closer its price movements come to being random). If you know the average daily change in price (and the average range of prices covered in a day) then you can calculate the probability of the price lying between certain values on the day or days following your starting day. 9. The real “strength” of the two sigma confidence interval is not well defined. I have done some of the necessary analysis and the answer is basically “it covers 95% of price movements, except when it doesn’t”. My two sigma confidence interval is very strong for a liquid, high volume stock within a confirmed trend. When the share closes outside this confidence interval it provides a good indication that there MAY be a change in the trend (other measures are needed to confirm the change in trend). Nothing can tell you with absolute certainty that a trend has changed on just one day’s data. [Though you probably don’t need any indicators if you see a sudden change of 20% in price] Once a trend change is confirmed the confidence interval continues to be “statistically” strong - it just no longer has any bearing on the future price movements of a stock. 10. The "strength" of my two sigma confidence interval is different for each share that is looked at. For most of the shares that I follow it seems to lie between 90% and 95% (for MIM and TLS it is only a little less than 95%, for OPS is considerably less than 90%). For all of the shares that I follow I assume that 95% holds - but I also know that I am overstating my confidence, to some extent. I do not know how strong it will be for shares that I have not tested it against (but provided the shares are liquid and trade well it should hold reasonably well). 11. As noted elsewhere in these notes - I have provided a separate spreadsheet, called “Confidence”, that will allow a user to determine the strength of a confidence interval and to determine whether a wider confidence interval will be better for their purposes. This is a slow and processor intensive task and so I have separated it from the main spreadsheet. 12. It is important to note that you can not draw a confidence limit for an illiquid share - the figures become meaningless. The less random the share movements are the harder it is to predict their limits. A lot of people seem to think that the opposite is true - that randomness means something is hard to predict. The limits of truly random events are quite easy, the price on any day is unpredictable but its limits will be well known. 13. Another important thing to note is that if I draw a 10 day forward projection of my two sigma confidence interval, based on today's data, it is likely be different from the same projection based on yesterday's data. Of course when I do it tomorrow it is very likely to be different from the results I got today. Maths and statistics do that to you. I often have no contact with the ASX for periods of weeks. I set my buy or sell orders and walk away. Once the orders are in I don’t continue analysing and re-analysing unless there are “fundamental” reasons (e.g. major announcements) to change my target price (whether it is a buy or a sell) I just leave it alone. 14. Unfortunately, the forward projections of the share price will tell me nothing interesting for the great majority of shares that I follow on the average day. If I have an open position on a share, on most days the projection will suggest that the price is not likely to reach my profit target or my stop loss within the planned time period. If I am just following the share, on most days, it will suggest that the share is unlikely to reach my target buying price. 15. A fundamental limitation with the method is that the results behave in an unexpected way when the share is nearing the bottom or top of a trend. At the turning point the projections will cross the peak or trough rather than follow them. This "flat" projection is only useful to confirm that the trend has really turned. For this reason the answer given by my “Confidence” macro is always going to be somewhat less than a real world answer. “Confidence” assumes that you still try applying the test during a confirmed trend change, while hopefully a real person would stop trying to apply the test in such a situation (the advantage of not be forced to blindly follow rules in the real world). 16. In order to address some of the problems that are experienced with this measure during trend changes I have included a short period linear projection on the 80 WD Chart. Any substantial divergence between the short and long period projections can indicate that the trend may be changing. 17. Overall I find these linear projections very useful for setting target prices for the buying and selling of shares and for confirming established price trends/trend breaks. Remember it is a guide only (as are all indicators). 18. As part of my analysis of the “strength” of my 95% I attempted to make use of the CONFIDENCE statistical function that is included with Excel. Unfortunately this function does not appear to work as documented. It produces confidence intervals that are a little less than half as wide as expected (covering less than 50% of the population of prices). As it does not appear to be possible to use the conventional function, I make use of volatility as a work around for a rigorously defined confidence interval which can not be readily calculated. M - Notes on the MACD Histogram 1. The MACD Histogram is a means of seeing trends in the movement of the price of shares. When the value of the MACD is positive the prices are trending up. When the MACD is negative the prices are trending down. Technically it is a trend following indicator and it confirms that a trend change has taken place - it can not be used to “predict” trend changes. 2. One way to use the MACD is to see it as a "Buy" signal when the histogram crosses from negative to positive and as a "Sell" signal when the histogram crosses from positive to negative. This is not a bad technique, but if you look at the actual prices paid you will find that you have paid a lot more than the lowest price and you have sold for a lot less than the highest price. In other words, you have "left a lot of money on the table". 3. A more sophisticated way to use the MACD is to make use of a signal line (drawn as a blue line on my graph). It is a "buy" signal when the MACD is negative and it crosses the blue line and it is a "sell" signal when the MACD is positive and it crosses the blue line. In order to make this clearer a red histogram of the difference between the MACD and its signal line is also plotted (if both the MACD and the MACD histogram are negative it is a buy signal, if they are both positive it is a sell signal). 4. If you use the MACD in this way, you will usually buy for (reasonably) close to the lowest price and sell for (reasonably) close to the highest price (you will still miss the sudden sharp peaks for a sell or the deep troughs for a low). 5. Unfortunately using the MACD this way does not work for all shares and even the shares that it works for it sometimes produces false signals (by that I mean, a "buy" signal followed by a decrease in price or a "sell" signal followed by a increase in price). 6. The fact that MACD is a good guide to your trading decisions on the majority of occasions will not make you feel any better when you follow it into what later proves to be a disappointing trading decision. It is important that you use indicators to assist you in making your trading decisions - you should not let them make your trading decisions for you. L - Notes on the RSI, MFI and volume divergences 1. In a continuing search for an indicator that tells me something useful about volume I have included the Money Flow Index (MFI). Various sources refer to this index (all noting that it is an analogue to the RSI), but none of the sources that I have found give an algorithm for calculating the MFI. I have not been able to find a source that plots MFI for stocks (this means that I can’t check whether my calculations agree with anybody else’s). Various sources give the same equation for “money flow” and so I have calculated “money flow” and then used the same intermediate steps as the RSI to produce an MFI. I have no way of knowing whether the indicator that I am calling the MFI is the same as that described in the literature, but my MFI does appear to be quite interesting and the RSI and MFI in tandem seem to be very useful. 2. The big question remains “How does one USE the MFI”. The answer appears to be that the MFI, by itself, is pretty much the same as the RSI and there is not much reason to choose one of these indicators over another (the patterns produced by the two indicators are often identical). We start to learn something interesting when we look at the RSI and MFI together and plot the difference between the two indicators. Sometimes the RSI will be strongly overbought (or oversold) but the MFI remains in the normal range - this indicates that the price action has taken place on low volume. Sometimes the picture will be reversed and the MFI will be overbought (or oversold) while the RSI remains unmoved, this means that relatively minor price action has taken place on large volume. 3. As noted above - how you interpret the divergences between the MFI and the RSI will be partly dependent upon how you interpret what the “smart money” is doing. If you view the “crowd” as ignorant sheep that follow the latest fashion then you want to know what the “smart money” is doing so that you can do the same. (Which would seem to make you a sheep as well - though perhaps a very perceptive sheep, significantly ahead of the flock). 4. Accordingly moves up or down on low volume can be seen as the “smart money” moving ahead of the “crowd”. If the “smart money” is wrong, the “crowd” won’t end up following. You will then be left stranded in what may be an extreme market position with no buyers. 5. If you view volume as an indicator of market sentiment then you will note that volume tends to dry up just before the market in a particular share tops or bottoms. You will then read moves on low volume as indicative of changes in sentiment (and take your money out before a downturn or put money in just as an upturn starts to happen). 6. Unfortunately market consolidations are also accompanied by “constrictions” in volume. The market pauses before continuing to rise (or fall) sharply. If you interpret a “constriction” on an uptrend as a top, you may take your money out before the market rockets higher (which is a bad thing). If you interpret a “constriction” on a downtrend as a bottom, you may put your money back into the market before a collapse (which is a very bad thing). 7. Constrictions in price and volume can be seen across a range of shares at various points in their price history. Shortly after my “conversion” to technical analysis I noted this (thinking that I was the first person to make the connection) and used it as my “golden indicator”. I noted the constriction in QAN, I bought it and it doubled. I noted the constriction in MIM, I bought it and it doubled. I then discovered a few cases where I noted the constriction just before a share price collapse. When I did my back testing I discovered that it is roughly a 50/50 indicator. It does presage big price movements, but roughly half will be up and half will be down (this can still be a money making proposition if you run tight stops and let your winners run - but it can be awfully hard on the nerves).
