APPENDIX A
FAQs

  • Why did you choose to become a market analyst instead of continuing to trade?
  • My goals involved working solely for myself, in whatever setting I liked, whatever hours I decided to juggle, doing varied and interesting work. I enjoy both solitary time writing and designing algorithms, and social/face time with clients during consulting assignments or teaching. It was also important for me to take time out for friends, exercise, and other pursuits, as I wished.
  • What is the most important aspect of trading?
  • In my view, there are two: wanting to trade because one finds the process interesting and enjoyable, and having enough money to do so without harming one's finances or the finances of others.
  • What is unique about your momentum indicators?
  • Relative to traditional momentum indicators, which generally evaluate current closes relative to past closes or high-low ranges, Kase indicators use entirely different mathematics, while displaying similarly to those indicators.
  • Kase momentum indicators are based on statistical measures of serial dependency, normalized for logarithmic volatility, and optimized for cycle length. The result is that Kase indicators succeed in triggering signals prior to about 80 percent of significant turns, while traditional indicators signal only about 55 percent.
  • It doesn't appear that you publish the formulae for your indicators in either the video or workbook. Why is that?
  • I try to describe enough about my indicators, including the basic math, so that those who understand the concepts explained, and who have the programming capability, could, if they wanted, duplicate or at least come close to duplicating my work, for their personal use.
  • However, questioners are often thinking about indicators like the DMI/ADX, MACD, Stochastic, RSI, Parabolic SAR, and the like. These indicators can be written with just a few lines of code. For example, the RSI = 100 – (100/(1+ (average of N days' closes UP)/(average of N days' DOWN closes))), where UP close = C – C[1] when positive, and DOWN close when negative (See Appendix D.).
  • Kase indicators or Kase studies are designated as such because they display on charts similar to traditional indicators or studies. In reality, these displays represent thousands of lines of algorithmic code in C++, Java, and/or C#, including complex arrays and object-oriented programming, that have taken countless hours over more than 20 years to not only fine tune but also to interpret with second-level displays like momentum divergence along with the momentum indicator itself.
  • In the old days, many technicians made their living by designing trading systems for hedge funds using traditional indicators, running investment funds, or teaching others how to trade, gaining name recognition by publishing simple indicators. There's nothing wrong with that—before about 1990 or so, there just wasn't sufficient computing power for indicators requiring more than a few simple calculations to plot. Therefore, it didn't make any sense for technicians to keep formulae secret.
  • In the early 1990s, it took 15 minutes to open one chart with my indicators, even with simplified code to speed it up. Now, with much more complex features, the indicators come up as quickly as any others. So, the work I have done, not only to invent good ideas, but also to program them, is valuable in itself, and considered proprietary intellectual property. I make sure to offer my work on a sliding scale, so that even individual, retail traders might use my studies and indicators.
  • For momentum divergence, what's the difference between an incorrect comparison and a comparison that's correct, but is nondivergent?
  • All comparisons evaluate peaks in price with peaks in momentum. Correct comparisons require equal or higher peaks in price in rising markets, and equal or lower peaks in price in falling markets, with associated momentum peaks. If the associated momentum peaks also make higher highs, and lower lows, while the comparison was correct, the formation is nondivergent. Conversely, comparing rising highs in a falling market, falling lows in a rising market, falling highs in a rising market, or rising lows in a falling market are all incorrect.
  • Why wait for a confirmation bar?
  • Momentum indicators use price and histogram peaks to generate signals. A peak, by definition, is surrounded on each side by shorter bars. Thus, it is impossible to tell that a price level or histogram bar is a peak until it is followed by a less extreme bar, for example a high price followed by a lower price, and vice versa, or a positive histogram peak followed by a less positive peak, and vice versa. Thus divergences and peaks are always identified no sooner than one bar after a peak.
  • Additionally, price peaks and histogram peaks don't always match. Usually, Kase uses a default of two or three bars for a tolerance where one peak can be before or after the other. Thus signals don't complete until one bar after the latter peak.
  • What is the correct “tolerance” to use when identifying momentum divergence?
  • When programming divergence, I find a tolerance of about two or three is good. The looser the tolerance, the more false positives, and the stricter the tolerance, the more turns will be missed. But there's really no correct answer. Under special circumstances large tolerances make sense. For example, very rarely oscillators can have very rounded histograms where the actual peak leans towards the right portion of the formation. In these cases, whether or not to consider a potential divergence as valid is a judgment call. On the reverse, markets that are trending in a choppy manner (I call these hybrid markets) can generate too many divergences with generous tolerances. In these markets, one might ignore divergences outside of a zero or one tolerance.
  • Why not sell right after a bearish divergence or buy after a bullish divergence?
  • The odds of a 3.6 standard deviation turn (based on double TrueRange, DTR) are roughly 33 to 45 percent. For example, with a Kase PeakOscillator divergence, if an average DTR reversal hasn't yet been hit, then the odds are 75 percent to hit one standard deviation over the mean DTR reversal, but only 33 percent to hit the 3.6 standard deviation turn. One way to profit from this is to drop down to a lower bar length and time into the trade there, then scale up if profitable.
  • What is the best time frame or bar length to use?
  • The best answer is that it depends. What is your time horizon? Are you managing a portfolio longer term, day trading, or something in between like holding a trade from a few days to a week or so? Most traders who ask this question are day trading or trading short term. Day traders, that is, those not holding a position overnight, should just find the level at which there is activity within each bar, and not flat lines for example, and visually ensure that there are enough bars in the day to reasonably trade in and out. We normally suggest no less than three-minute bars, or their equivalents, as the lowest time frame.
  • Once I choose a bar length, how do I determine the risk?
  • Once you've got a bar length, plot the Kase DevStops and look at the risk at Dev3, or KaseX, and then read the higher of the two risk amounts that display to the right. If you don't have either, then calculate the value equal to the average TrueRange plus four times the standard deviation of the TrueRange as a risk estimate.
  • What if the risk seems too high or too low for me?
  • If the risk seems too high for you, even at very low bar lengths, then you simply cannot trade that market. If the risk seems low, that might mean that the bar length is too short to overcome slippage and commissions; try a longer bar length. Day traders: keep in mind that the bars need to be short enough for you to enter and exit within a day. Intraday traders, holding trades at least for a few days: enter intraday and perhaps scale into a longer-term daily position. A bar length that yields about eight bars per day is suggested. Typically, one would start with about 34 to 55 bars per day, move up to 13 to 21, then when the trade becomes solidly profitable, up to 8 or maybe even as little as 3 to 5 bars.
  • What's your opinion on using time, tick volume, Kase Bars, or candlesticks?
  • The range of, and market activity contained within time bars are highly variable, especially for markets that trade overnight or have quiet periods. Tick volume bars are more regular. During quiet periods, it takes much longer to build a bar, and during very active periods, less.
  • Kase Bars are equal TrueRange bars. These bars have approximately equivalent ranges on each bar. Thus, the benefits of regular bars are taken to the limit allowable by the actual data.
  • Do I have to use multiple bar lengths when trading?
  • There's no overseer enforcing a particular trading style. While it's generally a good idea to use multiple time frames for reasons explained elsewhere, it's not always practical when trading very short bar lengths. For example, if you were to choose to trade a five-minute bar or equivalent as the normal chart you would be monitoring after scaling up; it wouldn't make sense to use a one- or two-minute chart to scale from. In this case, you would only use the one five-minute chart.
  • How do I know if there's enough room in the trade to profit when I scale down?
  • This is how I'd answer the question using Kase DevStops or the KaseX stops. These stops are associated with statistical probabilities. If the average reversal hasn't yet been hit (Kase's “warning line”), there's a 75 percent chance to hit the Dev1 stop. If the difference between the current price and the stop is sufficiently large, then I'd take the trade. But again, immediately drop down to the lower bar length where potentially short-lived trades would be more manageable, and use regular trading methods on that chart.
  • Keep in mind that odds increase for the next stop, should sequential stops be hit. Therefore, once Dev1 is hit, Dev3's odds increase from 33 to 44 percent.
  • Are stops really all that important?
  • Yes.
  • Traders don't enter trades expecting to lose money, so even if one is reluctant to place tight stops on a trade, there must be a point at which, realistically, one must admit the trade must be abandoned. Determine this point ahead of time, before the trade is taken, and stick to it as a matter of discipline. The same holds true even if a trade is profitable. Traders are often reluctant to exit profitable trades, as optimism is frequently blind to the reality that the run is over.
  • Placing a stop allows one to hold trades overnight without undue concern. If a market turns swiftly, having a stop in place might allow for better execution versus the time it might take to place a stop after the turn. Finally, real life happens. One might have a trade running and encounter a distraction or emergency, and fail to deal properly with the position. Having an “emergency” stop in place will help avoid further disaster in those cases.
  • What should I do if a candlestick confirmation point I'm using for a stop isn't hit?
  • The answer is simple, if you are long, stay long, and if short, stay short, relaxing if there's a close over a bearish or under a bullish pattern. This takes advantage of the false positive nature of candlestick patterns. If the pattern is indeed taking place at a turning point, you will be properly stopped out. If not, you will simply stay in the trade.
  • Should I wait for second long/short signals before entering a trade?
  • Keep in mind the “second-signal” police are not watching you trade, so you can do whatever you want! However, here's the point to taking second signals. Most formations have at least two impulse waves, such as ABC, where A and C are the two impulse waves, or a five-wave pattern, where there are three impulse waves, I, III, and V. However, in a given time frame a three- or five-wave pattern might condense (the wave pattern is only visible on a shorter bar-length chart) so that on your chart it's a one-wave correction. So the idea is to wait for the second impulse wave to make sure there will be one. If you have a special reason to believe that the trade you are taking is an exception—go for it. But, a much better strategy for trade acceleration is drop down to a shorter bar length to get in faster on its second signal.
  • Why is forecasting a market important?
  • Forecasting, or using a forecast, is like drawing, or at least using, a map. It's not that one cannot get from point A to B without a map, but knowing the probable path and likely signs for which to watch are very helpful.
  • What makes your forecasting techniques unique?
  • There are a couple of unique elements in my forecasting techniques, such as the use of Kase DevStops and candlestick confirmations and completions. Overall, though, because I use many techniques designed by others, such as wave projections, pattern analysis, and moving averages, what makes up my forecasting methodology—the recipe, so to speak? The answer is the programmatic manner in which I choose waves, calculate targets using a wide range of techniques, and determine confluence is unique.
  • How should I handle oscillating or choppy markets?
  • I'd suggest staying on shorter-term charts that are more sensitive to the oscillations, trade fewer contracts or shares, and wait until you have a solid profit buffer before scaling up.
  • How should I trade choppy markets?
  • Choppy markets are often corrective markets forming complex patterns, like flags, pennants, diamonds, and the like. These patterns are erratic, and it is never a good idea to trade them with heavy volume or in the time frames in which they are significant. If you want to be “in” a particular market, drop down to a lower bar-length. One bar length's oscillation may be another's short-term trend. (Longer bar lengths also can help, as an oscillation might disappear/condense into an impulse wave, but the risks are greater in the longer term.) So the key is that in any choppy market with a given bar length, regardless of how the indicator is used, the duration of the oscillation has to be sufficient to enter and exit successfully. Changing bar length is the answer to lengthening the duration in terms of numbers of bars or burying the oscillations in larger magnitude waves.
  • How can indicators help identify choppy markets?
  • The DMI, if whipsawing back and forth, forms a necklace effect, which is a sign of an oscillating market. A low ADX value, especially if the ADX is both low and declining, is also indicative of this. Any momentum indicator that has a zero line can indicate an oscillating market if it is oscillating around that line, especially if the values of the indicator are low.
  • How do momentum exit signals work in choppy markets?
  • Momentum exit signals work at least as well, if not better, in choppy markets as in normal markets. If you think about it, choppy markets oscillate, reversing up and then down, while normal markets do so less often. Thus, an indication that a turn is imminent is more likely to result in a reversal in a choppy market than in a market that oscillates less often.
  • Do you back-test your indicators?
  • Many of my indicators are back-tested for behavioral statistics, but not back-tested per se as part of a black box system. For example, I have tested the statistical performance of the DevStops—if a particular stop is hit or there's been a close beyond it, what the odds are of hitting or closing beyond further stops. Also, I've tested the reliability of the Kase PeakOscillator and KaseCD relative to both how often turns take place following signals, and how often turns were preceded by signals. I've done similar work for the Stochastic, MACD, and RSI.
  • What are black box or automated trading systems?
  • Automated trading systems consist of algorithm combinations where buy, sell, and exit signals are generated automatically. Black box systems are automated trading systems in which, for the most part, the indicators and rules employed are undisclosed, or mostly undisclosed.
  • What's your opinion on black box or automated trading systems?
  • There's nothing wrong with automated trading systems per se, with a few very serious provisos.
  • First, any back-testing must be done on real data, or to the extent that normalized data is used, tested for sensitivity to various methods of normalization. Typical high-low range bars, as opposed to Kase's Xrange bars, create bars with exact high-low ranges, filling in any “missing data” by inserting fake data. So, for example, if one sets a range target to 10 cents, and the security issues ticks of, say $2.01, $2.08, $2.15, a range bar will print a bar with $2.01 as the low and open and $2.11 as the high and close, with the next bar's open $2.11 even though there weren't any data between $2.08 and $2.15. Because of this phenomenon, high-low range bars that use fake data cannot be used in back-testing since the data never existed.
  • In back-testing futures data, it's common to normalize for rollover gaps. However, various methods of normalizing change results. For example, a method that tended to make prices higher would make win and loss amounts in absolute dollars higher and vice versa. So gains or losses could be magnified or lessened for this reason. This must be taken into consideration, especially if the system being tested has a low percent wins, but a high win-to-loss ratio.
  • Second, back-testing must be done over a sufficiently long historical period, covering a range of market conditions, testing at least a dozen or preferably more uncorrelated securities. For daily bars, for me, this would be at least 10 years or life of the security, whichever is shorter. The key is that any test should not only span trending markets, but choppy, erratic, oscillating periods as well. Additionally, tests should reflect times when markets might have gapped, making entry points much worse than the close upon which a signal was generated. Typically, Kase uses a minimum of 75,000 data points per study.
  • For intraday back-tests, even if one were to use dozens of securities over thousands of bars, given the short-term nature of the bars, the market conditions tested might have been only favorable. For example, there could have been a prolonged bull market lasting three years, in which case, a 25,000 10-minute bar test might have only captured mostly favorable conditions. So with short-term bars, both diversification of the test securities and sampling of varied data sets over a historically long period are needed.
  • Third, it's necessary to make sure that any results are based on realistic execution costs and trading conditions—not just reflecting brokerage fees, but also the difference between the time at which a signal is generated and the price at which the actual execution takes place. If, for example, a stop is hit in a gap, the exit could not take place until the following bar opens.
  • From this point, an initial step is to look at the stats and perform some simple risk-of-ruin calculations, employing the formulae presented in Section 8 of the video. These formulae have a limitation in that they don't reflect variations in percent wins and win-to-loss ratios, but only use averages. It's very important to look at the statistics on how many runs of losses took place, the maximum drawdown, and so on. The key is that a system might have, say, a 55 percent win rate but over a given three-year period on three of the, say, 20 securities tested, could have been 10 percent.
  • A better way to get a more realistic view of risk is to run a Monte Carlo simulation employing not only average results, but also considering the risk of bad runs and outliers.
  • Finally, the key to being successful with automated systems often combines with the ability to trade diversified portfolios. This way, if you are trading, say, 20 uncorrelated securities with the system, some markets might be favorable, while others are not. One can trade with relatively modest back-test results and still do okay, as profitable trades are allowed to run, while the losers are closed. Remember, taking trades in parallel is much less risky than in series.
  • What I really am asking is this: “What's your opinion on retail traders trading relatively small portfolios, using black box or automated trading systems?”
  • If a trader chooses not to trade a portfolio large enough to diversify risk, the key is having enough risk capital to withstand rare but possible long strings of losses, or unexpected hits from large losses. It's important to remember that even with stops, large unexpected losses sometimes occur. Gaps can occur on surprising news or random events by orders of magnitude relative to your stops.
  • One way to mitigate losses is to day trade, but even in this case, I would only suggest trading a system with a rigorous risk management system or exits and stops.
  • Personally, I think the best way to trade is to learn how to do it, using any fixed-rule trading system as a guide only. If a person is an aspiring trader, and has absolutely no execution experience, sticking to an automated system just for long enough to learn how to execute in a disciplined manner might be advisable. However, I would only recommend this after the trader has paper traded for a significant period, and again, transitioning to one's own style after a while.
  • How do you use volume in your work?
  • I don't use volume in my work for a couple of reasons. I've always focused on futures. As a futures contract becomes more prompt, it becomes more active. There might be lower and higher volume days, sure, but overall the time to expiration overwhelms any volume considerations. So, that's one reason why I haven't focused on volume.
  • The other reason is that I've always used either tick volume bars, or range bars, which tend to correlate with volume. Thus adding volume becomes unnecessary for this reason as well.
  • Can you explain how to differentiate between an expanding wedge and a broadening top or bottom?
  • If the formation is sloping up in a down market, or down in an up market, it's most likely corrective, so most likely a bearish or bullish expanding wedge. Again, if the formation is taking place well above a prior low for a possible broadening bottom or below a prior high for a possible top, that classification is less likely to be incorrect. The formation is probably an expanding wedge. Alternatively if the formation is at the top or bottom of a prolonged trend, and doesn't exhibit a slope, or at least not much of a slope, the classification is probably for a broadening top or bottom. When in doubt, other factors can help determine market direction.
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset