Lincoln B. Fiske Jr.
Founder, TradingVisions Systems, Inc.
A follow-up interview with Lincoln Fiske, Founder of TradingVisions Systems, Inc., a registered Commodity Trading Advisor (NFA # 347050) located in Spokane, Washington. Diversification is one of the most important keys to successful trading, and Mr. Fiske, a well known and prolific system developer, is one of the first to recognize the need to package various trading systems for customers interested in diversification. ?Vista Portfolios?, launched in November, 2004, offers (1) multiple trading systems, (2) trading multiple markets, and (3) trading in multiple time frames. He?s also been doing some interesting work using walk-forward analysis.
We know that as one of the true students of the futures market you are always working on your trading systems to adapt them to the ever-changing market conditions, as well as finding ways to improve their overall performance. Describe any changes you have observed in the futures markets since our interview in January of 2006, and what you have done to try to adjust to those changes, like with your Axiom Index programs.
Thank you for your positive comments, John, and for inviting me to be interviewed.
Trading in the last 18 months has been difficult. I focus primarily on the indexes, and there volatility and trendiness have remained low, despite occasional spikes. This has meant more whipsawing and fewer good trading opportunities. I think that the markets in general have become complacent in this environment of low, stable interest rates. Low rates affect many markets, and stability wrings volatility out of them because people know what to expect. However, these periods happen in cycles, and since we?ve been in this one a long time, I expect that we will eventually move out of it into more profitable times.
From mid-July of 2006 up to late January of this year, the Russell and Midcap?my two favorite indexes?starting behaving unusually, with abrupt, brief moves, followed by abrupt, brief countermoves. This, of course, is deadly for swing systems, and Delphi and AXIOM were no exception. Because of the lack of market follow-through and the smaller wave patterns, in March of this year I made the first change in AXIOM Index since its release in April of 2004, and I halved the exits. This significantly improves hypothetical results for the last couple years, and it has had a positive impact so far this year.
As a result of these market changes, last December I began looking for systematic ways to allow systems to adapt to the evolutions that markets inevitably experience. In the past I have been very averse to making system changes because it?s so easy to overdo this, but I have become convinced that there is a better way to approach system design. I have worked these last six months developing a proprietary approach to walk-forward optimization (WFO), one with articulated, sensible, repeatable rules. There is not much literature available about WFO, so I?ve had to learn primarily through testing and observing. It?s been an exciting and very time-consuming quest, and I think it may prove to be some of my most innovative and significant work. The initial results have justified the time, and I?ll eventually be revising all of the TradingVisions systems using this optimization framework. I started with Delphi, and because the WFO analysis suggested changing a number of the entry and exit rules and parameters, I decided to make the changes distinct by renaming it Delphi II.
Tell us how your new ?Walk-forward? protocol has improved your Delphi system.
It would probably be good for me first to explain what walk-forward optimization (WFO) is, and then I?ll give an example of the profound impact it can have on performance.
Most trading system developers use the traditional form of backtest optimization, which involves taking all the available past data and running tests of system parameters. Once the best ones are found, these are then used in real-time trading. There are two critical problems with this approach: 1.) traditional hypothetical performance records are idealized, "in-sample" results that usually vary widely from real-time results, and 2.) traditional optimization has no articulated way to adapt to different markets or to changes within a market. Walk-forward optimization helps solve both of these problems.
Walk-forward optimization (WFO) starts with a period of data--3 years, for example--finds the best parameter values for that period, and applies them to the next period of data, the next year, for example. The 3-year period is called the study period, and the 1-year period is the application period. The study period results are idealized because we were testing to find the ideal, while the application period is realistic, yielding results that we could actually have achieved, and this is the key advantage of WFO: its hypothetical results are in essence equivalent to real-time results, and they are therefore a much more valid basis for measuring the potential of a system?s performance.
The next step in WFO is to walk the data forward by now including the first application year as a part of a new study period, and dropping the first year of the original study period (in this example we are therefore using a rolling 3-year study period). We test the new study period, find the best values, and apply them to the next year. The results from this new application year are added to the first one, and we now have a 2-year hypothetical track record. This process continues forward, adding each new application year?s results to our hypothetical track record, until we get to our current year, where we are trading the best values found in testing the last 3-year study period.
At the end of the walk-forward optimization routine, we have a powerful hypothetical--but realistic--performance record of several years, showing the results of a system that gradually adapts to a changing market. This is the strength of walk-forward optimization.
The improvements WFO has made in Delphi are most dramatic in market-change periods like 2006 and early 2007. For the e-mini Midcap, original Delphi made $700 in 2006 and is down $3,000 this year, through the end of June; with a max drawdown of $10,000 (this is using the TradeStation? report, with $50 slippage/commission). Delphi II would have made $9,080 in 2006 and $1,465 this year, with a max drawdown of $6,200. Keep in mind that these results for Delphi II are not idealized, backtest-optimized results, but are from walk-forward analysis, using the rules and parameter values determined from testing prior years, so this is a fair and compelling comparison with the out-of-sample results of original Delphi.
I firmly believe that WFO is the best way to evaluate a system?s potential.
You were one of the first system developers to recognize the need for a ?portfolio? approach to systems trading, and to actually offer various combinations of trading systems by introducing Vista Portfolios. Since it was first introduced in November 2004, how have traders benefited using your Vista Portfolios to ?Harness the Power of Diversification??
One of the most basic principles in investing is to diversify, and this applies even more to futures, where there?s greater leverage. It?s human nature to go with the ?hot? system, and this may work for a while, but in the long run, it makes a lot more sense to choose a variety of systems, markets, and time-frames so that low-correlated equity curves help to cushion each other?s drawdowns. It?s the drawdowns or the risk factor of investments that typically causes a person to give up, so it?s very important to minimize risk. I?ve tried to do this with Vista Portfolios, which are combinations of the systems TradingVisions offers and the markets they trade.
Diversification works, and I?ve experienced this in my own accounts. You can see a simple example by looking at the 2005 and 2006 results that Striker reports for AXIOM Index and Impetus. In 2005, AXIOM swing trading the e-mini Midcap made 50% ($20,000 account, $20 commission). Impetus, which day trades the e-mini Russell, lost 10% ($7,500 account, $20 commission). I had many new clients who wanted to trade AXIOM EM in early 2006, and very few who wanted to trade Impetus ER. By the end of 2006, AXIOM EM was down 34% and Impetus was up 44%. Anyone trading both systems would have had a small profit for the year, rather than a loss trading AXIOM EM. Naturally we look to past performance, and it?s counter-intuitive to go with a recent ?loser,? but as long as a system is performing within its norms, it often pays to consider adding it to your portfolio so that it will be there when the equity curve swings back up.
You mentioned Impetus, which is now one of our longest-traded and most successful systems. Can you tell us a little about it?
I?ve been trading Impetus on the e-mini Russell since November, 2003, in my personal account, and it?s done very well there. Very few day trading systems have been able to weather the changes of the last few years, and Impetus is one of the survivors. It?s a dependable performer, with about 60% winners and a max drawdown of about $2,500, so it?s ideal for small accounts or for providing diversification in larger accounts. I also found that by changing the one main parameter slightly, Impetus does very well on the full-size S&P.
Is there anything in the ?What?s New? department that you care to share with our reader?s today?
I?ll continue to do walk-forward analysis, which will consume the next few months. I also have three new systems on the back burner that I hope to work on after that.
Thanks for the opportunity to catch up, John!
The preceding information not guaranteed to be accurate by Striker Securities, Inc. and does not constitute a solicitation of any kind. For further information please e-mail or call Dan Neenan 800-669-8838 / 312-987-0043.