Interviw of thomas demark.. where he soke abt his system at length...
Over the course of several conversations, DeMark explained the genesis of his analysis approach, his predisposition toward countertrend trading and the process of developing objective trading tools.
AT: Do you think you have a natural inclination to be countertrend, or was that simply an outgrowth of your early work experiences?
TD: I tended to look at things that way because I was always working in an institutional environment. In that kind of situation you’re dealing with size positions and you want to go the other way [sell into uptrends or buy into downtrends] to get your trades off efficiently.
AT: So you were never someone who focused on the longer-term trend and looked to enter on small corrections or pullbacks?
TD: Pullbacks still represent exhaustion, so that’s a viable technique. But you don’t want to be a typical trend-follower because of the price vacuums and gaps, which result in slippage. Trading commodities helped me a lot, because it made me realize that if you’re trying to catch a trend, you have slippage all the time. But if you go against the trend, slippage goes your way — it works in your favor.
AT: Do you think your techniques are applicable across markets and time frames — say, stocks and futures, and intraday bars as well as daily or weekly bars?
TD: Yes, all time frames and all markets. I don’t believe in developing a system or indicator for one market.
AT: You haven’t found that some techniques only work in certain situations or markets?
TD: No, I haven’t seen that, but I’ve never gone in that direction. I think of those kinds of things as optimized.
AT: Was your approach to break indicators open, so to speak, to see how they work and then try to improve on them?
TD: Yes, I did that with everything. There are certain concepts that have value, but I think people sometimes apply them incorrectly.
Amarket can basically do two things: It can trend up or down, or it can move sideways. But you can’t use the same indicators or techniques for both situations. Indicators like oscillators work in sideways markets but not in trending markets, and trend indicators like moving averages don’t work in sideways markets.
So I tried to find a different way to apply moving averages to address the reality of this situation. I don’t use traditional moving averages, but I use the concept of moving averages. For example, I’ll start the process when there’s a high lower than the prior 12 highs — that’s when I know the market is in a downtrend; the opposite would be true for an uptrend.
At that point, I’ll begin calculating a five-day moving average of the highs. If the market closes above the moving average and opens above that close the next day, that constitutes an upside breakout. But the moving average is only active four days, unless the market makes another high that’s lower than the prior 12 highs. Basically, the market has to provide evidence it’s in a trend — a high lower than the prior 12 highs or a low higher than the prior 12 lows — before I’ll apply the moving average. And when that prerequisite is no longer in effect, the moving average is cancelled.
AT: Do you have a preference toward a particular time frame or trade length — short-term, intermediate, long-term?
TD: Well, I don’t think it’s useful to think of things that way. Rather than thinking in temporal terms, I think it’s better to think of things in terms of percentage moves. For example, short-term might be a 5- to 10-percent move, intermediate might be a 10- to 25-percent move and long-term might be a greater than 25- percent move. In some cases, a trade might meet a long-term projection in a single day. If you accomplish your objective, you should get out of your trade, no matter how long it took.
AT: One of the early indicators you developed when you were at NNIS helped identify likely buyout candidates. How did that come about?
TD: I created a lot of models to measure buying pressure and selling pressure. The indicator was an outgrowth of an attempt to improve on the accumulation-distribution tools being used at the time. As it turned out, the indicator was finding buyout candidates before they were announced.
People were using close-to-close calculations, multiplied by volume, to calculate price-volume indices like on-balance volume. The problem with an indicator like that is you didn’t know when it would break out. Also, you couldn’t relate it from one stock or commodity to another.
The indicators I was developing were all volume weighted, but instead of a conventional close-to-close approach, I was comparing the close to the open and relating it to the high and low of the day. So, for example, if a stock opened on its low and closed on its high, you knew it was all buying pressure. The approach was to take the close minus the open, divided by the high minus the low, multiplied by the volume [(C-O)/(H-L)*V].
Now, that basic formula provided a cumulative index, but it wouldn’t tell you when a stock would rally or decline if you got a divergence; it also wouldn’t tell you which stock was better than another stock.
So I took the buying pressure divided by buying pressure plus selling pressure, which told me what percentage of the activity was buying pressure. Then I calculated the rate-of-change over different time periods — say, a 13-period, an 89- period and a 144-period calculation. This was the TD Pressure Index.
In the process of putting together this indicator, I noticed something interesting, which I hadn’t expected: Markets make their lows not because of buying coming into the market, but because of selling leaving the market.
That’s important. You can see the dissipation of selling as a stock goes lower — the change in the number of shares bought vs. the number of shares sold. Say you start out with 20 shares bought vs. 40 shares sold. As price moves lower you’ll see 18 shares bought vs. 20 sold, and then right at the low you’ll see 16 shares bought for 8 shares sold. The selling dries up. At tops you’ll see just the opposite: It’s not that people are selling at the highs, it’s that the buying evaporates, so by default, prices come down.
By taking the rate-of-change of the ratio of buying pressure divided by the combination of buying and selling pressure I was able to identify buyouts when the indicator got to a high extreme. There were five or six stocks out of a total of 32 I found over a three-year period where I went to management — we were managing their pensions in some cases — and told them, “Look, we can tell by this indicator that you’re going to be bought out.” One company told us, “There’s no way — we know where every share is in the U.S.” As it turns out, they got bought by a foreign buyer. They used to call me the Grim Reaper.
AT: Do you ever do a top-down kind of analysis, where you first look at sectors and then move down to analyze the most tradable stocks in that sector?
TD: Yes, you can look at the signals in terms of a sector index, for example, and then look for setups in the individual stocks.
AT: Have you done any research regarding market tendencies on certain days of the week, or leading up to certain times of the year?
TD: I haven’t really done much work there. I think there’s something to Mondays, because they often represent a premeditated move in the market. A gap on a Monday is significant because it gives you an indication of direction, especially if the gap holds. On Mondays people have had the weekend to think about things, and the major institutions have meetings before the opening. If they make a decision on a particular stock that has longer-term implications, it can cause a stock to gap open.
AT: Do you believe in using any kind of discretion in trading, or do you favor completely mechanical approaches? Do you see other people who are successful — perhaps using your tools — who blend in discretion?
TD: There are essentially three ways of looking at the market. Most people operate on the first level, which is to look at a chart and guess. There’s no consistency. The second way is to use indicators; there’s some subjectivity involved, but at least you have a roadmap. The third level is to use indicators in a very systematic approach.
Ninety-nine percent of people operate on the first level, three-quarters of a percent operate on the second level and the last quarter-percent operate on the third level. My feeling is that the typical chartist is a chart artist. I want to be a chart scientist. When I was developing techniques, I wanted everything to be mechanical.
But having said that, I know traders who use my techniques on a discretionary basis — they use them as confirmations. There has to be some discretion involved. Fundamentals ultimately dictate long-term moves and if you go against them, you’re going to be wrong.
AT: What are some simple techniques people can use to identify exhaustion points in the market?
TD: I have one approach called TD Camouflage that works especially well short-term. It’s based on being able to see “hidden” buying or selling pressure as market bottoms or tops are forming.
When they talk about the market being up today or down today, all quote vendors, newspapers and other financial media report price change relative to the previous day’s close. But what’s more important is price relative to the current day’s open.
A camouflage buy indication occurs when you get a close below the previous day’s close, but above the current day’s open. For example, if a market closed down relative to yesterday’s close, all the media would report that it was a down day. But if the market is up from the open, there’s really buying coming in. You can extend the approach by making sure today’s low is less than the low of two or three days ago. The reverse occurs at highs. If you look at the S&P futures, most of the turning points have TD Camouflage signals (see Figure 1, below).
[Tom DeMark Camouflaged Price Action]
If you look at July 11, the S&Ps closed down for the day. People think, “The market was down, it’s not going to turn up.” Well, it also closed above the open that day, and it turned the next day. Under the guise of a down market, the evidence of accumulation was being camouflaged. Now look at the high on July 19. It closed above the previous day’s close, below the opening and the high was greater than the high two days earlier. Again, that was the turning point.
It’s really perfect for short-term trades. If you’re a very short-term trader, you can use this to get in on the close and get out the next opening. The probability is very high for that kind of trade.
AT: How would you advise people to approach technical trading so they can identify these kinds of relationships and develop strategies?
TD: Well, you can start out by reading a good, basic book on technical analysis — I think Darrell Jobman’s book [Handbook of Technical Analysis: A Comprehensive Guide to Analytical Methods, Trading Systems and Technical Indicators (1995, McGraw-Hill)] is good — but you really have to study price action first hand.
You also want to turn off the news. There’s always an excuse for what happens in the market. You always hear things like IBM went up because earnings were good, or it went down because earnings were good, but they weren’t as good as expected. It’s stupid.
AT: Do you find it useful to treat the long and short sides of the market differently?
TD: No, I don’t think you can do that. My approaches are symmetrical.
AT: Don’t you see a difference in the way rallies and sell-offs behave?
TD: Seventy-five percent of the time the market will be in a trading range and any oscillator will work. It trends up 15 to 18 percent of the time and down 8 to 10 percent of the time. The reason is that buying is a cumulative process and the analysts become more bullish as the market goes up and people buy more on margin, but selling is a one-decision event. That’s why markets go down more quickly than they rise.
AT: But since that makes a difference in how markets move, don’t you think that calls for a difference in designing indicators and strategies?
TD: Well, you have a point, but I don’t think so.
AT: Can you describe the trendlinebased projection technique you used to make the interest rate call that caught Greenspan’s eye.
TD: Take the lowest low beneath the most recent trendline and calculate the difference between it and the trendline, and add that amount to the breakout point.
AT How did you make trendlines mechanical?
TD: For a down trendline — a TD Supply line — you take the most recent high preceded and succeeded by a certain number of lower highs and connect it to the next most recent high preceded and succeeded by the same number of lower highs. You can adjust the number of bars to make it longer or shorter term. Level 1 would be a high preceded and succeeded by one lower high, a Level 10 line would be a high preceded and succeeded by 10 lower highs.
You always connect two points only, and connect the two most recent according to these rules. Most people are accustomed to starting at the left side of the chart and connecting the most distant point and connecting it to the nearest point. But by connecting to the two most recent pivot points, you’re able to keep adjusting to the market.
Then you can use qualifiers to determine the effectiveness of the breakout. If a qualifier is there, it will probably breakout successfully. If a qualifier isn’t there, an intraday breakout will probably fail. Failed breakouts like that can be good one-day trades: You fade the unqualified breakout and follow the position with a stop.
AT: What are the qualifications you use for breakouts?
TD: There are a few. I’ll give you a couple of the basic ones. First, in terms of upside breakouts, there should be a down close the day before the breakout (see Figure 2, below). I’ve also experimented with a close less than the open.
[Tom DeMark Confirming Breakouts]
The second qualifier is an open above the trendline and above the previous day’s close, because in a situation where you might have a very steep down trendline the market could open above the trendline, but not above the previous day’s close, and trade one or two ticks higher. The move above the previous day’s close confirms the move. Also, for these confirmations, the open can never be the high — for an upside breakout.
AT: Are these rules the same for horizontal support and resistance and trendlines?
TD: Yes.
AT: How would you describe the logic behind these concepts?
TD: In the case of the first confirmation technique, if everyone’s piling into the market right before a breakout, there won’t be any money to propel it after the breakout. You’ll have an edge if you require a down close the bar before an upside breakout. If the previous close is down, there will be skepticism when it does break out, because most people have been getting short. People are creatures of habit. If the market has been moving down, they’re not going to cover or reverse until the close, so you’ll see a lot of vacillation around the breakout level. What often happens is that you’ll see capitulation in the last hour before the close and the market [will close] really strong when these people finally bail out of their positions.
I used to get burned by breakouts just like everyone else. I’d see a strong day the day before a breakout and think, “Ahh, it’s going to follow through.” So I’d buy the breakout and, invariably, it would close down on the day. So I had to figure out why, and I thought, “Well, if I’m buying on the previous day, or on the breakout, everyone else probably has already, too.” So, I decided I had to look for a down close the day before.
When there’s an up close before the breakout, the breakout will probably fail because everybody is already in the market because they’re all looking ahead to the market breaking out the next day. If everybody is already in, how can it go higher? It’s all psychology and common sense.
AT: Given that this technique is so reliant on what the closing price is indicating, would you use it on intraday bars?
TD: No, it’s really for daily bars. The close doesn’t have the same significance on intraday price bars — you could create seven-minute bars, for instance, and where a seven-minute bar closes is really irrelevant.
AT: Have you found indicators — that is, mathematical formulas — or price patterns to be more useful in trading?
TD: Price patterns. When I first started in the business, there were moving averages and trendlines, and that was about it. And then technical analysis exploded and ran the whole gamut.
At one point everyone [when I was at Tudor Trading] was working on neural networks and artificial intelligence, blah, blah, blah. They created hundreds of systems, and the best ones were the same three or four that I started with, despite all the research.
AT: What was the basis of those systems?
TD: Price patterns and qualifying moves, like we’ve been talking about. The basic concept is finding market exhaustion, which is the opposite of what most people do. Ninety-nine percent of commodity trading advisors, for example, are trend-followers.
AT: But some of them still make money. Don’t you think it’s possible to trade trends successfully?
TD: Sure. But my feeling is that there’s so much competition to get in on a trend you’re going to pay up for a buy, and then go through a drawdown immediately. I’d rather avoid that.
AT: What do you consider your best short-term indicator or trading technique?
TD: Sequential, applied to intraday charts.
AT: Sequential can be a bit confusing to people who aren’t familiar with it. Can you give us a brief overview?
TD: Again, you’re trying to find exhaustion points. First you need a setup — something that tells you whether the environment is right for a low-risk buy or sell. In the case of a Sequential buy setup, you need nine consecutive closes less than the close four days earlier. The market could turn at that point.
However, if it continues to go lower, you go into the “countdown“ phase: You take the close of day 9, and every subsequent day, and compare them to the lows two days earlier. When you have 13 closes less than or equal to the low two days earlier, it’s a low-risk buy point.
In an uptrend, where you’re finding a low-risk sell point, you need nine consecutive closes greater than the close four days earlier. There are a few variations and qualifications, but that’s the basic concept. It works on intraday data, too. I know traders who use it on oneminute bars. (See “Sequential signals”, for examples of Sequential.)
AT: When did you develop Sequential?
TD: In the mid-1970s.
AT: How long did it take you to develop it?
TD: Oh, maybe from 1973 to 1978. That was a very fertile period for me.
AT: Was that a process of studying printed charts?
TD: Yes. I developed almost all my techniques by hand.
AT: Do you think that perhaps having to do things by hand — rather than having the software on your desktop do it for you — helped you, in that you developed a unique approach and understood your techniques down to the smallest detail?
TD: Definitely. The real laboratory was my trading. If I didn’t trade, I wouldn’t have been as intimately involved in the market. I was losing money, and I had to develop things — necessity is the mother of invention. And I didn’t have much money back then. I would tell my wife, “This is my tuition.” I just kept pressing as much as I could. The result of pressing a bad position was that I acquired knowledge.
So, it was a twofold learning process: Examining charts — something you were forced to do because you didn’t have computers — and second, losing money. Those were the two catalysts.
If I’d had then the kind of computing power available now, I think I wouldn’t have had the same kind of investment I had in the whole process. I wouldn’t have been as closely connected to things.
And apurva when u strt a new thread .. then u can copy paste this article there also.... it will bring u many " THANKS"... hypocrate
Chintan786
