The following chart shows NZDUSD overlayed on AUDUSD. AUDUSD in the main chart (with the data shown as candlesticks), and NZDUSD is the overlay.

AUDUSD and NZDUSD (different scales)

Visually one might see that initially NZDUSD is “above” AUDUSD. Later on they cross one another and then NZDUSD becomes “below” AUDUSD. You might try to make somes conclusion out of these correlated pairs, but it is hard to make a good conclusion. You might think that since they cross, they will separate out again with NZDUSD being higher, but that is not the case as NZDUSD becomes “lower” as you continue with the chart. Later NZDUSD indeed goes up, crosses AUDUSD and indeed becomes higher again; so you might think “see, it works!”

But what is happening here is simple optical illution. In reality there isn’t such crossing at all. Or if the chart is presented this way, such crossings do not mean much. In many/most software, when two charts (or two series of data) are presented together, different scales are used in order to save space for the chart.

In this particular case, you can see the chart for AUDUSD goes from 0.9000 to 1.1500 (the scale is shown on the right hand side). The scale for NZDUSD (shown on the left) goes from 0.7000 to 0.8250. The total range for AUDUSD is 1.1500 – 0.9000 = 0.2500, but the total range for NZDUSD is only 0.8250 – 0.7000 = 0.1250. Since there are 5 grids on the chart, you can think of each grid represent a change of 0.025 for AUDUSD but only 0.0125 for NZDUSD.

In other words, if both AUDUSD and NZDUSD have a change of 25 pips, the AUDUSD chart will move up or down 1 block whereas the NZDUSD will move 2 blocks! That is basically what happened between 2/8 and 3/18/2011. AUDUSD moved down about 25 pips (I am estimating just by looking at the chart), and so did NZD. But AUDUSD moved about a block down whereas NZDUSD moved 2 blocks down (by blocks I mean 2 grid lines) because of the difference in scales. That is why you see the “crossing” when in realizty both moved about the same pip distance.

If a chart is constructed with the same data but only one scale is used, then more space is needed to present the chart, but the optical confusion is removed. With the same data as those shown in the chart above, I generated the chart below using Excel.

AUDUSD and NZDUSD (on same scale)

You can see that there is now only one scale (shown on the left from 0.7000 to 1.1000). I forgot to label the chart when I created it, but series 1 (the top chart) is AUDUSD and series 2 (the lower one) is NZDUSD. Because we are using same scale, NZDUSD appears below AUDNZD because the values are lower. However, the prices are moving basically in the same direction for both series, and one would clearly see no crossing in this chart and would not make trading decision based on the “crossings” seen from the first chart!

I hope this presentation helps in preventing one from making wrong or uncessary decisions in trading.

We tried to use arbitage (GBPJPY/USDJPY) as a measure of divergence, but when we compared the arbitage with the underlying currency (i.e. we calculated the correlation between arbitage and GBPUSD), we found that using arbitage is basically the same as using the underlying currency, GBPUSD. There is really not much advantage, if at all, of using arbitage compared to using the correlating currency. Furthermore, there doesn’t seem to be much advantage of trading the correlated pairs (GBPJPY and USDJPY together) instead of trading in the correlating pair (GBPUSD).

In this article, we will examine other pairs and show the correlated between the arbitage and the correlating pair. It might be easier for you to follow this article if you first read part 1 of the article.

Without further delay, here is the table for some correlated pairs. Each row shows a correlated pairs. The first two columns are the pairs, and the 3rd column is the proposed underlying pair. Arbitage is defined as the ratio of the first pair to the second (e.g. AUDUSD/NZDUSD for the first row). The arbitage correlation is the correlation cofficient between arbitage and the underlying pair. We are trying to examine if arbitage is indeed related to the underlying pair (in a few cases it is not).

Arbitage Correlation Table

One adjustment has to be made on the calculation of arbitage. In a pair like EURUSD and USDCHF, since USD is on the opposite sides, arbitage is defined as be the product of EURUSD and USDCHF instead of the ratio. This is true for any pair that is correlated negatively.

As you can see from the table, all the correlations between arbitage and the underlying pair has coefficients greater than 0.999! That is almost perfect correlation. That means the divergence and convergence of the correlated pair are determined by the underlying currency.

Let’s take EURUSD and USDCHF as an example. The pair are correlating negatively. The movements are almost like mirror images of one another. That is pretty easy to understand. The movement is determined by the strength of US dollar. If USD is strengthening, then EURUSD will go down and USDCHF will go up. It is just as simple as that. If the value of Euros and Swiss Franc never change, the EURUSD and USDCHF will mirror one another perfectly (except in scale) depending on the strength of US dollar. However, the values of Euro and Franc will not stay still forever, so instead of perfect mirror effect, they will fluctuate, and you can see that fluctuation in EURCHF.

Since the fluctuation is in some way unpredictable, there is no way you can predict the divergence or convergence. If there is no sudden and drastic event happening, in the long run the correlated pair will correlate even after a period of divergence. But perhaps here is where the fallacy lies: correlated does not necessarily mean convergence! For a positively correlated pair, most of the time they move in the same direction. Occasionally they might move in the opposite direction so that their “gap” closes, and after a while they might move back in the same direction, but that does not mean that the gap will always open up again.

In other words, you can buy EURUSD and USDCHF simultaneously to form a hedged pair, but you do not know when the pair is going to “open up” or “close down” in prices. You can use EURCHF as a helper, but the predictability of EURCHF is just like any other currency. You can never tell for sure. As pointed out in part 1 and can be seen in the coefficients above, trading in correlated pair like this is basically the same as trading in EURCHF.

In the next article I will show examples from real charts to hopefully make things clearer.

But there were times where the divergence was so much that I thought I should cut the loss. Depending on how much you use your margin, you might even get a margin call due to the divergence. I was (and still am) and conservative investor/trader, so it never happened to me, but there were situations that caused me to be quite nervous. On a couple of occasions, the pair didn’t converge after a long time (like a month or more), so I had to cut the loss.

The fact that it could diverge for about a month before converging back suggested that something else was going on instead of the common simple explanation that because each currency was traded in a broad market that sometimes there was a brief period of time where the divergence occurred. A month, to me, is not a “brief period of time.”

In those times I was just starting with about Forex, so I set up a small account to learn. As you know, trading a small account has a significantly different psychological factor than trading a large account even when percent gain/loss is the same. When I started to trade in a bigger amount (what is considered big is quite subjective), I knew I couldn’t just do the feeling-based trading as I did with correlation, so I did a formal study. This article and a few more related to correlation is the result of the study (the figures are generated recently, however).

To be a little more specific, I was using USDJPY and GBPJPY pair. It is easily to understand that they correlate well: they both buy JPY, so the prices usually (but not always) move in the same direction.  At first I thought about using a simple arbitage to tell divergence. Arbitage can be expressed by difference or ratio. I could use GBPJPY/USDJPY, for example, and see how the numbers vary. I suspected that the number woulds grow up and down and if Istudied the pattern (long-term and short-term) I could probably see some peaks and valleys and made decisions based of those.

And so I did the calculation of arbitage as defined above, for daily, hourly and 5-min data. I used 300 close prices for each of them, i.e. 300 days, 300 hours (12.5 days), 300 5-mins closes (25 hours). There is no specific reasons for using the same amount of data from daily, hourly and 5-min charts. It is just easier for me to program my spreadsheets the same way for all of them.

Here are the arbitage values that I obtained:

• DAILY: min = 1.433; max = 1.652 (difference = 0.219)
• HOURLY: min = 1.620; max = 1.659 (difference = 0.039)
• 5-MIN: min = 1.645; max = 1.655 (difference = 0.010)

I was a little surprise when I saw the differences between the minimum and maximum values of the arbitage. I was expecting the difference to be smaller with daily data but larger with 5-min data. Afterall, it was how I had been told for a long time: that the data between correlated pairs correlated well in the long run but that we could find divergences in short run and make profit out of it. It looks like the opposite is true. In short-term, the data didn’t show that much difference  in movement, but in the long run they showed some divergence.

But the next thing that occurred to me was, “wait a minute, how does that arbitage ratio of GBPJPY / USDJPY really tell me?” If you think about it, each pair is a ratio. When we talk about GPBJPY being 136.48, for example, we are talking about the ratio of Japan Yen over British Pound. So the notation GBPJPY really means JPY/GBP. Likewise, USDJP means JPY/USD. This means that the arbitage that I was going to use, GBPJPY / USDJPY can be written as follows:

(GBPJPY/USDJPY) = (JPY/GBP) / (JPY/USD) = (JPY/GPB) * (USD/JPY) = USD/GBP

But USD/GBP is simply what we normally call GBPUSD!

This realization more or less confirmed what I had suspected in the past: that the divergence or convergence of GBPJPY and USDJPY simply depends on GBPUSD. Actually this makes a lot of sense. If conditions affecting GBPUSD do not change, then of course GBPJPY and USDJPY would move in the same direction since they are both holding JPY. If conditions affecting Japanese Yen do not change, on the other hand, then the fluctuation would depend on conditions affecting GBPUSD. This is just common sense!

But while it is mathematically corrected that arbitage (GBPJPY / USDJPY) is the same as GBPUSD, they are not exactly the same thing because all three are separate pairs that you can buy and sell independently. Maybe there does exist some time where they diverge and we have a window of opportunity. Afterall, that is what the school of correlation investing usually tells you. So the next logical step was to check how closely GBPUSD is related to the arbitage, defined as (GBPJPY / USDJPY).

And so I did the calculation, and the result I got really surprised me (or should I say not surprised me because it had been my suspicion all the time): that arbitage between GBPJPY and USDJPY is high correlated with GBPUSD. In all cases (daily, hourly and 5-min data), the correlations are higher than 0.9!

Here are the correlations between arbitage (GBPJPY / USDJPY) and GBPUSD:

• daily: 0.99999
• hourly: 0.98347
• 5-min: 0.93575

As you can see, the daily correlation is, for our practical purposes, perfect (1 being perfect). Hourly and 5-min correlation are also high. On 5-min intervals, the number is smaller, but a correlation coefficient above 0.9 is still pretty high. That means that if you want to check the divergence and make decision to enter or exit a trade of correlated pairs GBPJPY and USDJPY, you might as well just look at the GBPUSD chart and make decision based on it.

What do all these mean to you

• Correlated pairs sometimes diverge in their chart, but the divergence is not caused by them being sold in different markets (and give you an opportunity to profit). It is caused by the fluctuation of the currency that correlate them in the first place (in this case the the divergence or convergence of USDJPY and GBPJPY pair is caused by GBPUSD).
• If you are doing or trying to do correlation investing/trading, rather than trying to calculate arbitage (in this case GBPJPY/USBJPY), you can simply look at the chart of GBPUSD.
• But if you can trust that you can make correct decision based on the chart of GBPUSD, you might as well trade in GBPUSD. Why bother with trading the GPBJPY and USDJPY pair when the profit/loss is determined by GBPUSD? As we all know, however, it is impossible to always make correct prediction of future price movement based on historical data. This also means that you cannot always trust that a divergence will always be followed by a convergence (especially in the short term).
• Trading in correlated pairs doesn’t seem to offer much advantage, if at all, unless you are doing some kind of hedging and buy and sell each pair at different time, but that is really a concept/strategy in hedging rather that correlation.
• The short-term correlation between arbitage and GBPUSB is high (0.93575) but not perfect (1.0), so there might be some brief period of time where there is true opportunity to trade in the correlated pair (instead of simply trading with GBPUSD); but 0.93575 is pretty high, so those opportunities are rare. Unless you are using computerized trading, you might not even catch them.

The big question is whether the above statements hold true just for the GBPJPY and USDJPY pair or whether they are general truths. In the next article I am going to do the calculations of correlation between arbitage of correlated pairs and the currency that correlate the pairs. Hopefully that will help our understanding of specific correlated pairs.

1. The momentum indicator ROC can signal an overbought or oversold condition.
This has some truth in it. If the momentum has reached a maximum (or minimum values), it means the price has been increasing (or decreasing) rapidly in the past N days, and the speed of price increase itself is increasing. The fast increase (or decrease) might be due to some news, so the peaks (or valleys) are signalling us to watch for further action.

But there is no way we can tell from the ROC or chart alone what direction the price is going. If the price has been increasing rapidly in the past, it could mean an overbought condition has been reached (a common use of ROC) so the price might be coming down. Or it could mean the trend is going upward, so although there might be a pullback, the price will to continue to rise. Note that either way is possible with ROC peak.

The converse is also truth. It is often said that oversold condition can be detected by the minimum values of he indicator. The problem is the minimum values could mean oversold condition, but it could also mean a down trend is starting.

Take a look at the chart of QQQQ for the past 2  years. Our eyes could be fooled to look at the valleys and think that if we had bought every time a valley was formed, we would have been rich. But watch closely. Don’t forget the common use of ROC also suggests to sell with peaks. If you had sold every time a peak was formed, you would have been broke.

http://finance.yahoo.com/q/ta?s=qqq&t=2y&l=on&z=l&q=l&p=&a=p12,v&c=

The reason that going long works but not going short on this chart is because the trend was upward. Had it been a downtrend the opposite would have been true. You might say that’s why we use the oversold conditions in an uptrend. But you do not know the trend beforehand. It is not that the valley acts as oversold in an uptrend, it is just that in an uptrend there are always pullbacks so the ROC is simply showing the pullback.

Below is another chart with ROC, except this time the price is decreasing. There are plenty of valleys before June 10, and if we thought they mean oversold conditions and bought on those, we would have lost a lot. BP was having problem with the oil spill, so people were selling, so the price dropped rapidly and so we had valleys on the ROC chart. They weren’t oversold, however. The ROC is only telling that the price is decreasing. It cannot tell whether it is an oversold or not in real situation.

 http://finance.yahoo.com/q/ta?s=BP&t=6m&l=on&z=l&q=l&p=&a=p12,v&c=

On July 15, there is a peak. We could have said from the chart that it was an overbought and sold BP. If we did, we would lose because BP went up. The reason it went up was because they seemed to have contained the oil spill. At least the initial test showed some success.

I hope I have illustrated enough here (and through my previous articles on moving averages) that news and events determine future price, and the price will be reflected in the chart, but the chart by itself does not determine future.

In summary, the ROC peaks or valleys can signal a watch condition, but it cannot tell the direction accurately. Future direction is determined by what is causing such a price increase and how the market reacts to it.

2. Bullish signal when Momentum rises above 0 for ROC (or above 1 for RSI)
This is really a very weak use of the momentum indicator. Think about this. What does it mean by the ROC rises above zero. Let us translate the statement into English: ROC rising above 0 simply means prices have changed from decreasing to increasing (see my previous article), but does that alone mean a bull signal? Are you really convinced that when the price action has changed from decreasing to increasing, it means the market is bullish and it is a time to buy?

3. In an uptrend, look for oversold (in an uptrend, buy if ROC turns upwards when below zero).

The problem is we don’t know whether it is uptrend or downtrend. This is why we are trying to use any indicator, remember? How can we now assume there is an uptrend and then try to use the indicator?

On the other hand, if you can somehow tell that is uptrend, then it doesn’t matter whether you use ROC to determine your buy signal. You can say use percent pullback or Fibonacci retractment for your buy signal. As with any indicator, you won’t be accurate 100% but you will do pretty good because it is in an uptrend. But the problem is you don’t know whether it is indeed an uptrend. Frankly, if you know how to confirm an uptrend, you will be the richest person in the world even without the ROC or any other indicator.

Textbooks and internets are full of statements like “in an uptrend, use the ROC for oversold condition”  or the converse of it. Yes, they sound good, but only if you can tell whether you are in an uptrend or not.

Summary for ROC momentum oscillator/indicator:

• Zero crossing doesn’t mean much.
• Peaks or valleys often indicate some news/events that caused rapid price increase/decrease.
• Peaks and signal can indicate “watch” signal, but other information is needed to make decision.
• The ROC (or RSI) oscillator cannot successfully tell an overbought or oversold condition.

In other words, we can use the RSI and start our process of research. We need to look at what has happened that caused the price increase/decrease and interprete the news if available. We need to also look at the overall market situation.

In this article we are going to look at the momentum oscillators. My belief has always been that in order to use an indicator/oscillator properly, you need to truly understand what it is and how it is computed and what it means practically (and that is different from how to use it magically). You cannot use any indicator as a magic tool as in buying or selling if the momentum is above or below a certain value.  There is no such magic number. Each tool must be used with proper understanding.

Definition and Calculation of Momentum

Momentum is a simple oscillator. It comes in two forms: difference and ratio. In the difference form the oscillator is often know as Rate Of Change (ROC). It is calculated by subtracting the price of N periods ago (for example, 7 days ago) from the last closing price.

The term rate-of-change (ROC) is actually better than momentum. The calculation of momentum shows that it is simply the velocity, speed or literally rate of change of the price of the security. In physics momentum is a product of velocity and mass. A truck hitting you at 50 mph is going to have an impact different than a cotton ball hitting you at the same speed, and that is because the mass of a truck is much higher than that of a cotton ball, and so the momentums of the two are different. The momentum known as ROC is simply a speed and so it is misleading to think of it as a momentum of the price movement of a security.

In the form of ratio, the oscillator is know as Relative Strength Index (RSI). The RSI computes momentum as the ratio of current closing price to close of N period ago. Because the computation methods are different, the figures for ROC and RSI are different, but they can be used the same way. In this article I will use Relative Strength Index (RSI) in the discussion.

Interpretation of Momentum

positive or negative values of momentum: since the ROC is the difference between latest closing price and the price N periods ago, it follows that if the ROC is positive, the latest close is higher than the price N periods ago. If ROC is negative, the latest price is lower. If the latest close is 30, and the close 5 days ago is 25, then the ROC is 5 (30 – 25) if we are talking about 5-day ROC. If the close 5 days ago is 36, the ROC would be -6 (30 – 36).

increasing or decreasing of momentum: if you see on a chart that the ROC is increasing, it means the rate itself is increasing. It does NOT necessarily mean that the price is increasing. It could mean the price has been increasing and is increasing faster now, but it could also mean the price has been decreasing but is decreasing slower now. Conversely, if the ROC is decreasing, it could mean the price has been decreasing and is decreasing faster now, or it could mean the price has been increasing but is increasing slower now.

flat ROC: if the price has been increasing but increasing at a steady pace, the ROC will be zero (or more likely flat than true 0 in real situation). Consider the situation where the price moving is 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, etc. If you do a 5-day ROC, the price different would be 5, 5, 5, 5, 5, etc (56-51, 57-52, 58-53, etc.) , so the ROC is flat. This is the part that is often overlooked or misunderstood. A flat ROC does not mean that the price is flat. It means the rate of change is flat.

zero crossing: if the ROC crosses the zero line from one side to the other, it means the price increase or decrease has change direction. Remembering that ROC above zero means price increasing and ROC below zero means price decreasing, it follows that when ROC cross from above zero to below zero, it means the price increase has stopped and now the price is decreasing. The converse is true for crossing from the other direction. If you look at a chart it might not be that obvious because most chart don’t use a 1-day ROC. In a 7-day ROC, for example, we are looking at the increase or decrease relative to the price 7 days ago.

Figure 1 shows the ROC (shown on the same chart) with increasing prices. Note that all the ROC values are positive because the prices are increasing. Toward the left the ROC is relatively flat because the price is increasing at a steady rate. Toward the right the ROC is moving faster because there is more price change (from steady pace to slowing down); but the ROC is going down although the price is still going up. This is ROC is a measurement of speed. At the very right, the price increase has slowed considerably, thus the ROC is approaching zero.

Fig. 1 5-day ROC with price going up

Figure 2 is the reverse of fig. 1. The price is going down and so the ROC is in the negative range.

Fig. 2 5-day ROC with prices going down

Fig. 3 ROC Crossing Zero

In the next article, we are going to look at some common uses of the momentum oscillator and discuss the strength of those uses.

Figure 1 is the chart of the price, 5-day and 15-day moving averages when the price is going up. The closing price is the blue line, and the 5-day and 15-day moving averages are the red and green lines, respectively. The first thing we note is that the moving average lines are below the price line. This is not surprising because moving averages are averages of past data, so if the price is increasing, the moving average will also be increasing but in a slower pace because it is lagging behind. You can also see that the 15-day moving average is lagging even more and so the 15-day moving average is below both the price and the 5-day moving average.

Fig. 1 Price, 5-day and 15-day moving averages with price going up

The next chart is a chart when the price is going down (figure 2). Quite naturally the opposite from above is true: the moving averages are above the price line, and the 15-day MA is above both the price and the 5-day MA.

Fig. 2 Price, 5-day and 15-day moving averages with price going down

Fig. 3 Price, 5-day and 15-day moving averages with price changing direction

And that crossover is the famous (should have been and should be called the infamous) use of single moving average. What is happening here is since the price HAS CHANGED direction, the moving average line FOLLOWS and so it crosses the price line. But what is always or very often suggested is that since the moving average crosses the price line, the price must be changing direction. That is a logical fallacy!

The exact same thing is happening with the 2 moving averages. When the price is increasing, the 5-day MA (moving average) is above the 15-day MA. When the price is decreasing, the 5-day MA will become below the 15-day MA. That means at some point the two average lines must cross, and here is another famous (should have been and should be called the infamous) use of two moving average.

The same fallacy is occuring here as well. The changing direction of the price HAS CAUSED the 2 moving averages to crossover. What is often incorrectly suggested is, however, that the crossover of the two moving average suggests that the price will change direction. The crossover does tell that the price HAS changed direction, but whether the trend will continue cannot be predicted by the crossover.

(My apology for the poor chart. I am still looking for an online source where I can capture a chart and use it in my blog without worrying about copyright problem. I still haven’t found it and so I have to download financial data and use Excel to plot the charts myself.)

HOW GOOD IS MOVING-AVERAGE CROSSOVER

If you are just starting in charting, you might be reading a lot in technical analysis and charting. You have probably been convinced to read MA crossovers, so you eyes would naturally look at point B and see how well it is in “predicting” the down trend.

But in order to be fair, you really ought to look at point A too where the crossover is suggesting an up trend. It later turned out to be a false signal. If you had the stock purchased right after point A and you followed the crossover signal and sold right after point B, you would have lost about 9% of your investment in about one and a half months.

But that is not too bad, you said. Sure we have false signals from time to time, but when the signals are right as in points B and C, they are really good.

Afterall, if you sold short at point B and then bought at point C, you could have potentially made about 45% return! If you follow the moving average strictly and only buy after point C, you would have made 28%, but that is 28% in less than 3 months. That is certainly not bad, is it?

So the true question is: had the moving average crossover at point B successfully predicted the down trend that follows? I might have said this a thousand times before. Well, actually I haven’t, but the answer is simple and loud and clear: NO.

STOCK PRICES AND NEWS

I guess it is time to reveal what stock this is that I am talking about. It is BP (BP PLC). Since it has been on headline news almost every day for a while, I figured everyone must be somewhat familiar with it, so it would be good to be used as an example.

Following the chart, point B occurred on May 3 (Monday), but notice that the price has dropped a week earlier. What happened was the Deepwater Horizon drilling rig explosion happened on April 20. Eleven workers died and 17 others injured. At first the impact was still unclear. On April 24, a report from Homeland Security said the problem had “no near-term impact to regional or national crude oil or natural gas supplies.”

But on Apr 25, US coast guard remote underwater cameras report the well is leaking 1,000 barrels of crude oil per day (bpd). BP’s share felt 2% the next day. Very soon it was realized that this was a huge environmental disaster affecting the US and the world. Quite obviously it was (and is) going to cost BP a huge amount of money, so the stock price dropped.

A couple of months after the explosion, it seemed like BP still did not know what to do (or at least the news made it sound so). There had been a lot of blame on BP, on the government, and the president of the US; but frankly, I think the problem was just so big that nobody knew what to do yet (at least at that time). Well, if you had a big problem and you did not even know how to tackle it, people lost confidence in you, so the price continued to drop.

On June 1, US launches a criminal investigation into the oil spill. The price dropped with a gap from previous week. On July 11 to 13, BP removed the old cap and replaced it with new cap. Initial test showed promise. The price went up a little. On July 15, BP said the leak had been stopped by capping the gushing oil wellhead. That is when you see a spike in the price.

I could have given more details, but the point that I want to say it these events are the major reasons for price increase or decrease. And when the price increases or decreases, the chart is affected included how the short-term and long-term moving averages look. It is definitely not the other way around.

WHAT COMES FIRST

On put it in another way: the explosion caused the price drop. The realization of the size of the problem caused it to drop more and so the price crossed the moving average from above. The realization that there was not yet a known solution caused to drop even more and faster and hence the down trend and the crossover of long-term and short-term moving average. It is just that simple, isn’t it?

Or do you really think the crossover of the price with the first moving average caused the explosion? And then the crossing of the two moving average caused the undersea camera to detect the leaking of 1,000 barrels of crude oil per day and hence the down trend? Or BP found that they didn’t have a good solution because the moving average crossover predicted that the price would go down?

BACK TO OUR PREDICTION GAME

Now let us go back to the original question. Is BP stock price going to go up or down? What do you think the moving average crossover is telling you? How much effect will this crossover have on the future price of BP?

As of today, we know the capping the wellhead was not the final fix for the problem. It was part of a test. Other things need to be done to permanently fix the problem. The plan is to drill relief wells around the original ruptured oil well and to permanently plug the leak. BP is facing legal issues and fines as well.

If the repair effort is going well and the leak seems to be permanently plugged, then the price might go up because some investors might see the current price as a bargain. If the repair process is delayed (e.g. by yesterday’s storm) or if unforeseen but not-too-serious problems are encountered, the price might continue to fluctuate. If the new cap burst and so capping doesn’t seem to be able to contain the problem, the price will go down.

BP is also facing a lot of legal issues and fines. If we hear good news on these issues, then the price might go up. It will go down if it is the other way around.

So how is the recent moving average crossover going to affect the price? Quite frankly? Nothing. It will have zero effect on BP’s price. The moving averages and their crossover look this way because of what has happened in the past, not because of what will happen in the future.

The reason I dislike most of the usages of moving average is simply because the logic involved is often reversed: if things were going nicely (like if the price had been going up/down steadily with normal fluctuations), then the moving average chart would look nice too. However, we are often led to believe that when the charts look good, prices are going to be good and predictable as well. That is the logical fallacy that I described in my previous posts.

However, there is some use of longer-term moving average that might make some sense if properly used. The one I am thinking of is the use of 2 moving averages and their crossover.

Crossover of Two Moving Average

 A caution must be made first. With the contradiction and fallacy mentioned in my other posts, I am not saying that all of the sudden the longer-term moving averages become so useful in predicting future. My view is that moving averages are never good in predicting future as it is just a graphical view of what has happened in the past.

But that is exactly how we are going to cautiously use it: to see the past. Moving averages are averages of past data, so they do a good job in smoothing out small variations / fluctuation of prices in the past. If we use them that way, we are using precisely how it is meant to be used.

Because of its smoothing feature, when you look at a 50-day moving averages, for example, and it is changing direction, it is often true that the longer-term prices (if you consider 50 days to be longer term, of course) have reached a certain low or high and is now changing direction. If this is confirmed by a shorter-term moving average (e.g. 10-day moving average), it means prices have changed direction in the longer-term and have now been confirmed by the shorter-term prices.

But that alone should not trigger a buy or sell signal. It is just a signal for you to start paying attention. The current economic condition, recent news about the company or the funds, the global economy or political situation and many other events will affect the price. The crossover tells you that you can at least start watching the particular security and make considerations along with other factors.

And that is why I always tell others that you need to know exactly what an indicator is telling you. If moving average chart is a chart telling you something about the past, sure it is able to tell you something about past. This might sound redundant but I can’t overemphasize the important of this principle: to know exactly what a particular indicator is, how it is constructed / calculated and what it means. Sometimes you need some math to be able to understand it, but you certainly cannot just trust somebody else’s writing. I believe many authors have different motivations / agenda when they write.

Now let’s take a look at a chart with 2 moving averages (DIA, 1 year, 15- and 30-day moving averages). If you look at points A and B, you can see that at or near those points, the prices have indeed reached some high or lows in the past. To be more precise, it has actually past its low or high, but moving average is a lagging indicator, and we all know that.

You should also pay attention to points C and D where the crossover “signals” failed. Again, ultimately it is the fundamental issue or news that drive prices.

So here it is. At least I am presenting some good use of moving averages. The principle is worth mentioning again: if you use an indicator for what it really is, then it is ok. Don’t use an indicator that is specifically designed to plot past data (as moving average does) to predict future.

DIA 1 year Daily

In the next post, we will try to predict future with moving averages! You are all invited.

In this post, we will take another look at some common usages of moving average:

Usage 1
Truth: In an uptrend, the moving average acts as a support.
Fallacy: If the price touches the moving average from above, it is a signal to buy since it has reached the support.

Fig. 1 IBM 6-month daily with 15-day moving average

But what really happened is the chart looks like this because the price is (or we should say emphatically it has been) going like that. Some explanation is needed here:

When the price has been going up, the moving average is below the price because moving average is a lagging indicator. This is seen between point A and just before point B. Then the price moves downward a little, but since the moving average is slower than the price, the moving average is not catching up with the price downward movement as fast. The result is that the price will get near the moving average or touching it (point B). Now since this is an uptrend, the price moves up again (usually for other economic reasons rather than because the price touches the moving average). Since the price is moving up again, the moving average resume the position of being below the price (between B and C).

In other words, the movement of the price causes the moving average line to look like what it does. It is not the touching of the price and the moving average that is causing the price to go up the second time.

The same thing can be said about the opposite, that is, in a down trend, the moving average acts as a resistance. Since it is just a mirror of the original statement, it will not be discussed here.

Usage 2

Truth: If the price has been moving up and is now moving down, the price will cross the moving average from above.

Fallacy: When the price crosses the moving average from above, a sell signal is generated.

I deliberately choose the same chart to illustrate the point. Typical a chart like this is usually shown, except that now point E is highlighted. The chart looks very convincing especially when one is looking at point E and notices that the price moves downward after the crossover.

Fig. 2 IBM 6-month daily with 15-day moving average

But what really happened is that there is a crossover because the price has been going up but then turning down. It is not the crossover that is causing the price to go downward. We have covered this in my previous post, so I won’t go into detail here. What I would like to discuss is the contradiction between the first and second usages.

The Contradiction

If you haven’t noticed it, please look at the two common usages again. The first usage says that if the price touches the moving average, the price will go up. The second usage says that when it touches the moving average, the price will go down. So which one is true? In other words, if you are using moving average to generate buy or sell signals, when the price touches the moving average, should you buy or sell?

The fact is neither is right. Or if you so desire, you can also say both is right because the price is going to eventually move in one direction or another; but the price movement is not determined by the fact of the touching or crossing over at all.

And if you are consciously aware of this contradiction, and you look at charts that suggest usage 1 or usage 2, you will notice the contradiction quite clearly, very often in the same chart. At point F, the price is touching the moving average, and later the price go down (so they say the moving average acts as a resistance). At point G the same thing happens again. At point H, however, the same touch happens, but the price goes up afterward (but now they position and say the touch/cross means the trend is changing direction). Whatever the direction the price might take, they can say they are right. The problem is how do you know which right is right?

Point I is worth examining too. It is touching the moving average from above, so it is similar to point A, B, C and D on the previous chart. According to usage 1, the moving average would act as support and so the price would move up. It did not. It moved down. Point J is the opposite. The price meets the resistance and so should move down. It did not. It moved up.

There are plenty of contradictions in charts like this. It is just that your mind is led to see only certain ones that an author happens to be writing about. Now that you can consciously look for this kind of things, please examine different charts honestly. You will notice that when an author is suggesting and illustrating a point, you can see a lot of contradictions happening in the same chart that he or she is using to illustrate his/her point!

There are a few things that one must clearly be aware of when using moving averages:

Average by itself does not tell much.

Let’s say the closing prices of a particular stock for the past 5 days are 5.1, 5.2, 5.3, 5.4, and 5.5 (increasing in price), and another stock has the price of 5.5, 5.4, 5.3, 5.2, 5.1 (decreasing), yet another stock with 5.3, 5.3, 5.3, 5.3, 5.3 (unchanged). All of them would have the same moving average of 5.3. The average by itself does not tell you anything about the market condition.

Moving Average uses old/past data.

When one uses moving average, he is actually observing the change of the averages (usually graphically). If you are using a 10-day moving average and you are looking at how the averages increase or decrease in the past 10 days, you are actually looking at 20 days of old data.

Some people use 20-day and 50-day moving averages, which means effectively you are looking at data about 70 or more days ago. Frankly, in this fast-moving internet age, I don’t know if the price 70 days ago is going to have that much effect on current price. For example, in today’s news, Toyota is subpoenaed by a U.S. federal grand jury for documents related to problems in its vehicles. That fact and the eventual outcome of it are going to affect the stock price of Toyota more than its price 70 days ago.

Moving Average is a lagging indicator and must be used as such

There is a common logical fallacy involved when moving average is used. In logic the fallacy is called the fallacy of affirming the consequence, and it goes like this: if it rains, the ground is wet; so if the ground is wet, it must have rained. But that is not always true. It can be true but it cannot be argued from logic that if the ground is wet, it must have rained because there are other reasons why the ground can be wet. Another way of saying it is if A then B does NOT necessarily mean if B then A. Unfortunately that mistake is almost always committed when one uses moving average.

Example 1

fact: When the security is in an uptrend, the moving average is going upward, and the price is above the moving average.

fallacy: When the moving average is going upward, and the price is above the moving average, the security will go up.

The factual statement is correct simply because of the definition of moving average and so the moving average is lagging. But a common use of moving average is fallacy part. The price has been up in order for the condition to be true. The price can indeed increase further for thousands of reasons, but it can also decrease for thousands of other reasons, but it just cannot be concluded that since the moving average is below the price, the price is going to go up.

Example 2

fact: If the price has been in an uptrend and then the trend reverses, the price will cross the moving average.

fallacy: If the price has been in an uptrend then the price crossovers the moving average, the trend is reversing.

Again, the fact part is true by the definition of moving average so moving average is lagging. Because the closing price is going to go from above the moving average to below it, at one point it is going to cross over it (see the chart below).

But notice how another common use of moving average is based on the same logical fallacy: it is so frequently said that if the price line crosses the moving average line and become higher (or lower) than the moving average, it generates a buy (or sell) signal.

Usually a chart like shown above is presented (in this case, QQQQ from Feb. to Jul 2010, with its 25-day moving average) and the chart looks very convincing. Your eyes are led to see that the price line crossed the moving average and then the price seems to go down indeed; but what happened was actually the opposite! The price first stopped increasing, then decreased and so it eventually crossed the moving average. That is how the moving average line is constructed! The price drop was causing the crossover, not the crossover causing the price drop.

If you are not used to thinking this way, I know it may sound unconvincing, but think about it for a moment or two. In my future posts I will show more charts to illustrate this point.

PS: Your input and feedback would be greatly appreciated whether you agree with me or not.