Nice line-up of speakers

[DSteckler]

The line-up of speakers at the upcoming AAPTA (American Association of Professional Technical Analysts) conference this weekend is impressive. I won't be there Friday but will be there Saturday, and will provide this list with summaries of each speaker's topic.

Friday:

George Schade – The history of Relative Strength and a Current Sector Model

Ryan Harder, Rydex Investments – Using Technical Analysis in Asset Management

Mike Epstein, MIT – Behavioral Finance Musings

Keynote speaker: Larry McMillan, McMillan Analysis – Using And Trading VIX

Roundtable Session, Moderated by Richard Arms, Arms Advisory – Market Internals as Intraday and Daily timing signals and how they have changed over time. – This is an open forum in which all are encouraged to discuss their observations on how the use of market internals such as breadth and volume has changed, and any unique ways they can be used.

Bill Sharpe, Valern Investment Management, Chairperson IFTA, 2006

Saturday:

Larry Winer, Ned Davis Research – Robustness Testing in Relation to Data
Segmentation

Linda Raschke, LBRGroup – Torturing Data

Roundtable discussion – Long Term Cycles in Commodities and Energies

Peter Brandt – Trading with Classical Chart Patterns


I'm especially looking forward to hearing Peter Brandt since I incorporate chart patterns and price projections therefrom, e.g., triangles, H&S, etc., in swing trading.

[dmk112]

Where is this held?

[IIC]

Where is this held?

It's in Orlando http://www.aapta-us.org/

[DSteckler]

The first annual conference was held in Chandler, AZ, a suburb of Phoenix.

[DSteckler]

A few of the speakers are friends. George Schade is a helluva nice guy who is an attorney specializing in water rights; he is a market historian. Mike Epstein has been around since the abacus was new and is a hoot; we had lunch together last fall in one of the MIT faculty dining rooms and he had me and another friend of mine in stitches the entire time. And Linda has more energy than any three people combined!

[IIC]

A few of the speakers are friends. George Schade is a helluva nice guy who is an attorney specializing in water rights; he is a market historian. Mike Epstein has been around since the abacus was new and is a hoot; we had lunch together last fall in one of the MIT faculty dining rooms and he had me and another friend of mine in stitches the entire time. And Linda has more energy than any three people combined!

I've heard Linda speak a few times...she seems very sharp...As far as an attorney being a helluva a nice guy though...Hard for me to fathom...I'm not one to stereotype...But I don't like attorney's...Including my own...Doug

[DSteckler]

Gee, and I thought you liked me, Doug.

[IIC]

Gee, and I thought you liked me, Doug.

Well...I do...But are you a practicing lawyer???...That might make a difference...Doug

[lemonjello]

Chartistas

Just curious about how all these chart people got hooked onto the MM fundamentalist stocks.

Is it like the Great White shark and the remoras (shark suckers)? The shark does the swimming, the stalking and the killing and then the remoras get off to eat and congratulate themselves on their magnificent hunting ability.
















Just kidding.

[lemonjello]

Where's my laugh track?!?!!!

[DSteckler]

Well...I do...But are you a practicing lawyer???...That might make a difference...Doug

Not for a loooong time.

[DSteckler]

Here is a summary of the speaker's comments. I took those down as they spoke so they may come across as a bit disjointed. There was a Q&A session after each speaker finished. I didn't arrive until Friday evening but I understand the speakers were every bit as dynamic an interesting Friday as they were Saturday.

Larry Winer, Ned Davis Research

Robustness Testing in Relation to Data Segmentation



Discussion of segmentation of historic data in an attempt to improve objective, task-tested indicators and models, while limiting over-optimization.



Utilize cross validation techniques like in-sample, out-of-sample testing.



In-sample/out-of-sample. OOS used to provide feedback/validation and if desired to confirm or reject the selected model. Results should not be used to provide feedback to re-optimize the same model for newer parameters.



Use the last 10% - 25% of the data for the OOS period. This simulates most recent years as real-time data.



For intermediate and long term testing, use in sample and OOS testing periods. For example, use three years of data of for in-sample and two years for OOS.



For short-term model testing, one practice is to alternative between even and odd years for IS and OOS testing. This has the benefit of breaking up any long-term trends. This preserves the integrity of the IS/OOS concept, while possibly diminishing the true action of individual trades.



Periodic testing: If the majority of the gains were obtained in one or two periods, then the indicator/system would be suspect.



Regime testing: Are models influenced by environmental conditions (e.g., valuation components, external influences, commodity influences). Each regime needs to be defined. This testing is best used for indicator/model confirmation but not for final model development.



V-Fold Cross-Validation: Uses two distinct periods of data for the IS period (training and validation). Works well with various modeling techniques such as regressions analysis, ranking systems, and indicator models. Provides a way to select the best specifications for a model; it is not used for parameter optimization.



There are far too many combinations of testing to comfortably say that any one method should be employed.





Linda Raschke, LBRGroup, Inc

Torturing Data



Is there statistical significance to TA?



She looks at time of day tendency, price tendency, exits/entries, with price only. Looks to see if the models work in multiple time frames, multiple markets, etc.



Presentation is going to show that through quantitative testing, there is a small, significant edge to both simple rules as well as complex strategies.



Price behavior. Randomly entering a position can produce a profit 65% of the time. By applying simple filters, this can be breakeven or even slightly profitable. This data gives a benchmark for the systematic trader.



By adding a trend filter, random entries can be made into a profitable system. Only long entries are taken in an uptrend and only short entries in a downtrend. The profit target is 4 ATRs and the Stop Level is set at 3 ATRs. The system was 100% profitable with index futures but there are periods of strong drawdown. The system was profitable between 92% - 100% of the time in all markets except live cows and hogs, silver and natural gas.



A system can be considered better than random entries.



Volatility breakout or Channel breakout as an entry. Use an increase in range. There is a statistically significant positive expectation to the system. Works best when played for a small target or short-term period (1 – 2 days). Much of the edge of these systems get eaten up by friction (e.g., slippage).



It is folly to try and anticipate which breakouts are money makers.



Momentum Driven models. Momentum functions are based on a derivative of price. Increase in noise over smoothed functions, e.g., MAs. Smaller targets are more profitable than large targets. Time stops are better than fixed stops and time stops combined with fixed stops are optimal. Momentum driven models perform better with a weighting given to an additional use of time stops. Look for price highs to be confirmed by momentum highs.



Momentum filter + Trend filter. The win % rate stays the same, the number of trades is reduced, and the net profit may increase slightly. System profitability may decrease since fewer trades are taken.



Modeling extended runs. “Extended run”: a more persistent imbalance in the supply/demand relationship. This is characterized by an increase in market efficiency, decrease in noise, and thus lends itself to using trailing stops. Use a 5 period SMA. Seven closes on one side of the SMA are considered to be significant. Enter on the first close on the opposite side of the SMA and exit on the first close that is in the direction of the original trend. %Win = +70%; average loss about 2X average win; profitable in 90% - 95% of a basket of 22 markets in any given year.



The % win on entering on the first retracement trade after an extended run dropped off significantly after 18 consecutive closes on one side of the 5 SMA line.



Market structure as measured by swings. Pioneer in the area was Arthur Merrill. Waves eliminate noise. A new Up Wave starts when price moves N-ATRs up from the lowest low or lowest close over an N-look back period. A new Down Wave starts when price moves the same ATRs from the highest high or highest close over N-look back period.



Basic wave rules. Trade in the direction of the prevailing trend, playing for a fixed percentage.



There is no way to predict with statistical consistency how far a trend will move.



Hsu and Kahn tested 39,832 trading rules, contrarian rules and complex trading strategies. Conclusions: Profitable rules and strategies can be found more readily in “young” markets such as the RUT and Naz as opposed to “mature” markets like the Down and SP500. The best rules are all short term (two day) MA rules. Most profitable rules and strategies are based on filter rules and MA rules. No contrarian rule is significantly profitable.





Roundtable, moderated by Ian McAvity, Deliberations

Long term Cycles in Commodities and Energies



Cooper may have $1.50 as the low. Triple top breakout. GSCI Commodity Index broke out in 2004; rectangular base from 1974 - 2004. Gold/oil; median ratio 1971 – 1004 is 15 bbls oil = 1 oz. gold.



Dow/Gold ratio. 1900 – 2006 secular trend. Median ratio since 1968 is 9.2X. 18.20 oz. = DJ 11,347.



Scanning 144 years of SP500 data, great bull markets are an anomaly. Much more time spent in broad, multi-cycle ranges.



Gold vs. SP500. 1968 – 2006 weekly close, ratio was almost identical. Charts looked almost identical. Gold and gold stocks are two distinct asset classes with different trading characteristics. We’re in a very, very mature cycle for gold; it’s been 121 months since the TSE made its last cyclical low, the longest it’s been since 1968. He’s convinced gold will see $3,000/oz. within the next 10 years but it won’t be next week.



Dollar will probably go back to $1 - $1.02. Peaked in July 2001 and bottomed in Dec. 2004. The $1 - $1.02 is the retracement.



Expects rates to go a lot higher. Says Bernacke is a student of Fed history and will much more follow Volker than Greenspan. He expects the Fed to be much more aggressive under Bernacke than Greenspan was.



Gold/silver ratio since 1900. Historically, 15:1. Currently 45.8X. Silver top tend to occur in April and October.



(He talked very little about energies, spending most of his time discussing Fx and metals)



Peter Brand, Factor Trading Co.

Trading with Classic Chart Patterns



Every month he compares how he traded against his trading rules (signals). He grades himself based on how accurately he followed his rules, not by how much money he made.



Classical chart configurations: OHLC

Signals – Major: weekly patterns, 10+ weeks in duration

Major: continuation signals, 1 – 3 weeks in duration

Minor: Daily patterns, 4 – 10 weeks in duration

Specialty – Triple non-confirmation with gap; hooks; 4-day away



Entry – At breakout – intraday; breakout on close

Risk – Major – up to 2 ½% of assets

Major pyramid – 1% - 1 ½%

Minor - .7% - 1%



Protective stop placement – Last Week Rule; Last Day Rule (last day/week the price is within the pattern. That’s where he places the stop once he enters the position)

Profit positions – Target; Target-plus; Reversal patterns; moving average; target extensions



Do not trail stops – let system come all the way back down (the Law of Least Regret; he’d hate to miss out on a major move). Number of contracts based entirely on risk.



Wrong 65% of the time. Use of open orders/resting orders to enter and exit – goal is 100%. Enter and exit using stops.



Basic trading unit is $100,000. 35+ markets followed and traded.



Cardinal Trading Rules: Will this trading signal (win or lose) be clearly identifiable on the charts a year or two from now? Did I take every signal? Did I manage the trade correctly? Did I take undisciplined trades?



“Best Dressed List.” No question about weekly daily patterns; decisive breakout and run to target; minimum challenge of entry point.

[IIC]

Do these people actually trade...Or do they just talk about it???

[DSteckler]

Do these people actually trade...Or do they just talk about it???

ROTFL!

Larry Winer is a research quant for Ned Davis Research and previously, he was a portfolio manager for a mutual fund. Linda Raschke - go a Google search, Doug. She runs several hedge funds, among other things.

Iam McAvity co-manages a gold fund in Toronto which holds more gold than almost any other private institution in the world.

Peter trades over 35 commodity and futures markets on a daily basis.

Speakers Friday were George Schade, who spoke about the history of relative srength; Ryan Harder, a portfolio co-manager for Rydex; Mike Epstein, a full proffesor at MIT who was managing money before you were born; Larry McMillan (I'm sure you've hear of him); Richard Arms (ever hear of the TRIN index? It's also known as the Arms Index); and Bill Scharpe, who is a money manager.

[IIC]

ROTFL!

Larry Winer is a research quant for Ned Davis Research and previously, he was a portfolio manager for a mutual fund. Linda Raschke - go a Google search, Doug. She runs several hedge funds, among other things.

Iam McAvity co-manages a gold fund in Toronto which holds more gold than almost any other private institution in the world.

Peter trades over 35 commodity and futures markets on a daily basis.

Speakers Friday were George Schade, who spoke about the history of relative srength; Ryan Harder, a portfolio co-manager for Rydex; Mike Epstein, a full proffesor at MIT who was managing money before you were born; Larry McMillan (I'm sure you've hear of him); Richard Arms (ever hear of the TRIN index? It's also known as the Arms Index); and Bill Scharpe, who is a money manager.

Geez...I spend at least 45 hours a week researching stocks...55 hrs a week at my job...But it seems like Linda at $795 a pop to attend her seminars does better than me...But she does so many seminars...When does she find the time to run a hedge fund???

BTW...What is a Research Quant anyway???