The Tradescapes Professional Edition offers extended analysis options for the financial professional:

Referential Tradescape Procedures (6)
Sentimentscape Procedures (5)
Book-Half Reverse Pyramiding Procedure (1)
Timed Exits Procedure (1)
Backtest Engine Customizations

Each of the options added to the professional version are designed to solve specific problems in entity selection and signal design that are particularly difficult or onerous to to address with any certainty.

Referential Tradescapes – Surrogate or Referential Signaling

There are two principal reasons to employ a reference as the target or source entity for trading signals that will be applied to a different traded entity. The first is where a basket of entities is traded using a signal derived from a single entity.

Basket Trading

Each Financial entity has its own unique historical signature in terms of sweet spots or time horizons where it can be effectively signaled. Certain entities will favor a faster signaler, others a more patient one. There will be those with little lag tolerance, especially for the more basic signaling algorithms, and others that may be quite forgiving. The idea behind using a single entity to trade a basket of securities is that the entity upon which the signaling is based will be stable, consistent, and as minimally prone to adverse fat-tail events as possible. This is often done with sector indices, or the major securities that often lead a sector.

One trades a basket of securities looking for an edge over the simpler process of trading the master entity directly. The problem here is the unique signature of the different components to be placed in that basket. Even though they may be represented in the overall index or entity, they may signal in very different places were they to be traded independently. Referential tradescapes are a way to determine which entities belong in a basket based on their historical trading landscapes as derived from the intended surrogate or master entity. They not only assess which entities will trade in synch with the overall entity, they will also highlight those whose chaotic behaviors have too little lag tolerance to be usefully placed in such a basket.

The Referential Tradescape procedure in the professional version is invaluable for finding meaningful surrogates and viable candidates that work with a given surrogate. This procedure can take a process that would require days of work, and leave you with no certainty afterward, to a process that can literally be accomplished in minutes, and after the fact you will know you have done everything you can to have the best possible match between surrogates and traded instruments. The Referential Trading Signal Analysis procedure will evaluate your own real-world signaling algorithm, applying it to the reference, and using those signals to trade the candidate entity selected. The performance is plotted atop the referential tradescape.

The above referential tradescape panel has XLB directly signaled in the first panel, and then in the following panels we see XLB’s current components with at least 10 years data history and at least a 4% weight in the ETF, each traded using XLB as the source for the signaling. There are huge differences across the individual entities. Referential trendscapes are used to build baskets that are most likely to be successfully traded using a single surrogate.

Difficulty in Ordered Signaling

You may encounter instances where the entity that is being traded is too chaotic to work well with the algorithmic signaling systems you use for your trading. This is usually a product of fat-tail events that often work against any trade that may be underway. There are entities whose price movements to a new trading range occur abruptly. In such cases, algorithms are often fooled in terms of making unfavorable entries and exits. If you can find an entity that closely tracks the entity to be traded, without the chaotic movements, the additional lag tolerance may translate into a much more effective signaling, despite the additional ‘fuzziness’ the two entity system will add.  In practice, this is a very challenging problem to solve. How do you select a surrogate? How can you know it is actually accomplishing its purpose? The referential tradescapes reduce a very difficult problem into a manageable one, and there will be a much higher degree of confidence in any solution.

Sentimentscapes – Two Stage Sentiment Signaling

In a two-stage sentiment signal, a slower time horizon signal is used to determine the direction of trade that is allowed, but does not actually initiate a trade. The faster signaler then executes only those trades permitted. The slower time horizon signal can be drawn from the traded entity, a type of confirmation signaling at two different time horizons, or it can be drawn from a more global entity, such as a market or sector index that would be a proper reflection of the overall sentiment currently in play for a given entity.

Two-Stage Confirmation Signaling

You will at times find two different time horizons within a tradescape where each could potentially be traded. Both may have a strong performance and good lag tolerance, and the progressive tradescapes may show both as being viable across time. When this occurs, it need not be an either-or choice. You can use the traded entity as the target for generating a signal at both time horizons and then the slower signals can open up the long and short (or out of market) trading windows, and the faster signaler will signal the entries and exits. Without using sentimentscapes, the two-stage tradescape technology, this type of signaling is almost impossible to implement with any confidence.

Two-Stage Sentiment Signaling

This adds another dimension of complexity. The slower signal that sets the allowed direction of the trades must be specified not only for this greater time horizon but also with the entity from which that sentiment will be drawn. The entire universe of financial time series is available. An overall market or sector index may be the obvious choice, but the sentiment may be better reflected in a different entity that leads the one you are trading. There will thus be less lag in the slower signal that sets the bands of allowed direction. This is a problem that most would deem close to intractable. A signal designer could literally spend weeks designing and carrying out the different studies. The amount of information that would have to be processed can be immense, even unmanageable if most of the backtests are inconclusive. Confidence after the fact will be hard to come by. Sentimentscapes offer the means to build such a signaler in an astonishingly short time, and with the maximum of confidence in the design. Overfitting isn’t a factor.

BHP’s single stage directly signaled tradescape is in the first panel. The sentimentscapes that follow increase the time horizon of the second sentiment signal, here based on the Hong Kong index. Note how the reward-pain and the lag tolerance at a very fast time horizon is markedly improved with the slower sentiment signaling.

Two-Stage Referential-Sentiment Signaling

Certain entities will simply be very difficult to signal. As a financial professional there may be considerations where such a challenging entity must be placed under active trade management. There may be instances where a surrogate signaler can be used for the primary entry-exit transitions and a sentiment signaler with yet a different entity from the entity under trade can be used to represent the overall market sentiment. In this unique case, we have two signals, the two entities to which they are applied, and the third entity that is the one actually traded. In the past, we would have regarded this problem as fully intractable. The amount of effort needed to reach a successful signaler would be prohibitive, and confidence would be close to zero. In a single procedure, the referential sentimentscapes in the professional version make this possible.

The Advanced Trading Signal Analyses procedure in the professional edition will evaluate your own real-world compound signaling algorithm against a sentimentscape or referential-sentimentscape.

Book-Half Partial Exits (Reverse Pyramiding)

Professional traders sometimes employ a strategy of booking a portion of the profit within a trade when a fixed percentage gain is realized. This is often a very simple and effective form of reverse-pyramiding. Booking half of one’s profit at a fixed percentage gain locks in half the current return from the equity under trade and still allows a portion to continue in the trade. By having this function applied repeatedly within a long trend, the equity can be minimal at the point where the turn occurs. The return is diminished, but the pain, as defined by drawdowns or retracement can also be dramatically diminished.

The two panels above reflect 10 year EOD long tradescapes for SPY. The first has no book half algorithm and the second uses a 7% fixed book half level. In terms of reward-to-pain, as RRt, the lag tolerance improves as well the robustness across different time horizons.

The purpose of Book-Half tradescapes is to answer the question as to whether or not this type of reverse pyramiding can significantly improve reward-pain. If such occurs with the accurate signaling used within the tradescape, there is a fair chance it may be even more of value in real-world signals that have less accuracy in catching the turns. As one would intuit from the nature of this approach, entities that suffer chaotic behavior at the turns in the signal are those most likely to benefit.

Book-Time Exits

Professional traders sometimes employ a simple strategy of booking a trade at a fixed point in time if a stop does not otherwise occur. If an entity is particularly hard to signal in its exits, it is possible one can find a solution by exiting before the trade would on average typically turn. The principle is a basic one. There is a certain entry edge from the accuracy of the signaler’s entry transitions, and this will be at a maximum some count of bars into the future. By trading a timed-exit one can significantly avoid the lag issue. One often exits before the turn even begins to occur. This may be inelegant, and it may miss a portion of the return, but the strategy can self-correct and re-enter if the signal is still favorable after a fixed count of bars following an exit.

While these types of exits tend to lack robustness, they can sometimes be surprisingly effective. The above are 8yr EOD long tradescapes for GLD. The left panel has the normal exit signal and the one on the right books all trades at 50-days with no re-entry until a fresh entry signal occurs.

The purpose of Book-Time tradescapes is to answer the question as to whether or not this type of exit can effectively deal with entities that are otherwise onerous to signal for exits.

Customizable Backtest Engine

The professional version exposes the backtest engine behind the various tradescape analyses. The standard version uses a single equity model no trading costs, no stops, and simple signaling on the close. The professional version offers the following customizations:

Starting Equity Amount and Equity Model

The backtest engine can be customized for three different equity models:
Amt is Initial Equity-All Equity Invested in Each Trade
Amt is Fixed Equity Invested in Each Trade
Amt is Fixed Shares Invested in Each Trade

Trading Costs

A fixed Cost Per Trade and a Spread/Slippage (%) can be specified.

Trade Timing

This is also another way to account inefficiency in execution:
Trade Close – Value and Execution Taken Just Before Close
Trade Close at Open of Next Bar – Value at Close, Execution Next Bar’s Open
Trade Close at HL of Next Bar – Value at Close, Execution Midpoint of Next Bar

Trade Stops

A stop is set with the Risk % for Stop and a Cooling Period (bars) accompanies a stop out. There are six different types of stops:
Bar % Intrabar-An adverse % from Bar Open Signals Intrabar Stop
Static % Intrabar-An adverse % from Entry Price Signals Intrabar Stop
Trailing % Intrabar-An adverse % from HH or LL Since Entry Signals Intrabar Stop
Bar %-An adverse % from Bar Open to Bar Close Signals Exit at Close of Bar
Static %-An adverse % from Entry Price to Bar Close Signals Exit at Close of Bar
Trailing %-An adverse % from HH or LL Since Entry to Bar Close Signals Exit at Close of Bar

Trade Signaling

Signal is Based on Close of Bar
Signal is Based on HL of Bar
Signal is Based on HLC of Bar

Automatic EOD Portfolio Generation

Whenever incoming daily (EOD) data contains multiple entities, a equal weight portfolio of all entities (independent reference and sentiment entities excluded) can be automatically generated and made available to the backtest engine. In the various tradescape procedures that support EOD data, you may specifically select the portfolio, or for those options that generate a panel of different entities, the portfolio will be included as the last panel.

These backtest engine options offer the means to study the impact of these settings on the ideal signaler used in the tradescape system. In particular, we felt it a very useful approach for determining the most efficacious trade timing, stop levels and types, and signaling variable.

Overfitting and ‘Touching the Data’

Experienced professionals understand the reality of random luck and the dangers inherent in overfitting. If you process a thousand different referential entities against an energy company, there is a certainty you will find a candy company or a health-care entity or another unrelated entity that tracks almost perfectly. The features we have included in our professional edition of Tradescapes acknowledge the importance of ‘touching’ the data to be traded as minimally as possible in the process of signal design. We consider it essential. In our tradescapes paradigm, we ‘touch’ the data ‘one’ time whenever we perform a full-accuracy trading map or a panel of such maps. Each tradescape analysis is designed to represent what we regard as a single ‘touch’ of the data.

You touch the data once to determine the time horizon to trade and to see how good you must be with your signaler in terms of the lag tolerance you are granted for that entity. That is often as far as you need to go. You must still find or build the real-world signal that delivers that time horizon and lag fraction, but the design is over after a single touch of the data. The direct-signaled options in the standard edition manage exactly this.
For more challenging signaling, you may again touch the data each time you explore a surrogate. You may touch the data a second time to explore a sentiment signal’s time horizon when used in a composite signal. You must touch the data again for each sentiment entity, but an experienced professional is likely to know which entities can represent the overall sentiment of the entity under trade. You may touch the data once more to explore continuous leverage position sizing, or to see if the book-half strategy can improve reward-pain, or if a difficult to signal entity can be managed with timed exits, but in all cases, the count of touches will be small.

The tradescapes professional version offers advanced trading science, but we go out of our way to avoid the pitfalls of overfitting and overoptimization.