Questions about Entry/Exit Systems and Signals

Questions about Entry/Exit Systems

Why are my entry/exit systems not producing exit signals when the exit rules are true?

The rules associated with an entry/exit system are checked in the following order:

·      Enter Long

·      Enter Short

·      Exit Long

·      Exit Short

 

In other words, the exit rules will only be evaluated when the entry rules are not true.

 

The order of rule evaluation cannot be changed. However, you can add logic to your entry rules to check for relevant exit conditions and not enter in those situations.

 

So that you do not have to enter the same logic in the entry and exit rules, you can create a function definition with the logic for when to exit. Then, evaluate this function as part of both the entry and exit rules.

Questions about Optimal Signal Generation

Why did the optimal signal generator not generate any signals?

The optimal signal generator uses several parameters to determine when to generate entry and exit signals. Two of these parameters are the number of samples into the future that are available and the minimum profit per trade. If the minimum profit is set too high for the number of samples being examined, no signals will be generated. This can happen when the trading price is relative stable.

:    Example: If the minimum profit is set to 10% and the number of samples into the future is set to 5, the trading price will need to change by enough in a five day period to create a 10% profit after the entry and exit commissions are deducted. For a $1000 investment and $10 commission, the trading price would need to change by a little over 12% in five days.

Why did the optimal signal generator produce losing or low profit trades?

The primary thing to check here is to make sure that the signal analysis you are looking at is using the same trading style as the optimal signal field. The trading values for automatic signal analysis can be viewed from the Signal Analysis Options Dialog. Remember, there are both global settings and data-level settings for this analysis. If a data series has its own settings, they will override the global settings. The trading values for signal analysis fields can be viewed from the Modify Field Dialog for the signal analysis field. If the optimal signal is being analyzed using different trading values than it was analyzed for, it may produce losing or low profit trades.

 

Losing and low profit signals can also be produced for the analog and thresholded analog types of optimal signals. This occurs when the market appears to be a good opportunity for a short trade on one day and a good opportunity on a long trade soon afterwards. With the maximum profit type of optimal signal, the signal would wait for the peak or valley before changing position. However, with the analog and thresholded analog types of optimal signals, an opposing entry signal may be generated before the profit of the first signal is realized.

Why are there empty cells at the end of my optimal signal field?

Empty cells represent null values. Null values are typically produced when there is not sufficient data to calculate a value. This can be due to one of the inputs to a calculation or prediction being null, or a calculation or prediction requiring a previous value that is not yet available. In this case, an optimal signal field uses a specified number of future samples to generate an appropriate signal. If the required future samples are not available, the optimal signal will not be able to be generated.

Why did my prediction produce a higher gain than the optimal signal?

There are several situations in which an entry/exit system or a model of the optimal signal can produce higher gains than the optimal signal.

 

The primary reason that this will occur is that the optimal signal is not calculated at the end of the data series. Since the optimal signal generator uses a given period of future prices to determine the best position, it can not be calculated when those prices are not available. This causes the optimal signal to not change positions during that period, possibly resulting in losses or missed trading opportunities.

 

Another reason that the optimal signal can be beaten is the way it is calculated. The optimal signal is generated so that it only takes advantage of changes that produce a minimum gain. By default, this minimum gain is 5%. However, trades yielding lesser gains can be effectively used to greatly increase the overall gain. Because of this, models of the optimal signal can produce better gains than the original optimal signal. This is especially true since these models are optimized for maximum profit by default.

 

An additional reason is that the analog optimal signal is not an optimal signal that produces the maximum profit. The analog optimal signal will produce an entry signal as soon as it is possible to make the minimum profit. However, the maximum profit may have been gained if the trade was delayed. This is reflected in the strength of the analog signal. To analyze or chart the maximum profit signal, set the type of the optimal signal to maximum profit.

Ä    Note: The maximum profit optimal signal does not include strength information and, therefore, may be more difficult to predict.

Questions about Signal Analysis

Why is the By Bar analysis delaying executing my signals by one bar?

When a signal is generated, it typically requires the values on the current bar to be available. The earliest possible time that this signal could be executed would be during the next bar. (The exception to this is when daily bars are used and after hours trading is considered. Unfortunately, after hours trading does not have a bar associated with it.)

 

To account for this, signal analysis executes a generated signal at the start of the bar after it was generated. This being said, the price used for this trade can be set to be based on the opening price for that bar, the closing price from the bar that the signal was generated, or other values.

Why does the amount of equity at the end of the By Bar analysis not match the overall or By Trade analysis?

In order to simulate completed trades, the overall analysis and the By Trade analysis both close the last trade internally, causing an exit commission to be paid. The By Bar analysis does not pay this exit commission.

Why are my optimization targets not being reached even though I have a positive fitness value?

If you have set target values in a trading style and these targets are met, you will typically get a positive fitness value. Likewise, if you get a negative fitness value, one or more of the targets may not have been met. However, when you are looking at the signal analysis, you may encounter values that do not reach their targets even though a positive fitness has been reported. Why is this?

 

The primary thing to do is to compare the date range of the signal analysis to the date range used for the optimization. The fitness values were set based on the ability of the targets to be reached within a specific date range. The values for another date range may be significantly different.

 

Another thing to consider is that it is possible to get positive fitness values even when some targets are not reached. This is especially true for targets set for low importance, or targets that were narrowly missed relative to the overall expected range of values.

Why is the final gain (loss) on a trade larger than the maximum loss?

The maximum loss for each trade is set using the trading value. In other words, this is the worst loss you could receive by performing a normal exit from the trade without slippage. A larger loss may be incurred if the stop price is reached and this price is worse than any preceding trading prices.

Why does the signal analysis have a different number of trades from the charted entry/exit signal?

The analysis displayed on the Modify Field Dialog: Signal Analysis page may not match the entry/exit signals displayed in a Chart View or Spreadsheet View. Here are some common causes for this:

·      Trades Excluded by Analysis Date Range

The date range being analyzed may not include all of the data. This can cause entry signals that appear in a chart or spreadsheet to not be included in the analysis.
This date range is reported on the
Overview page as the Analysis Date Range and can be modified in two places. The default signal analysis options can be set from the Modify Options Dialog: Signal Analysis page. The signal analysis options for this data series can be overridden from the Signal Analysis Options Dialog.

·      Trades Excluded by Analysis Activity

The signal analysis may be only performing long or short trades. This can cause all long or short entry signals that appear in a chart or spreadsheet to not be included in the analysis.
The analysis activity can be set in two places. The default signal analysis options can be set from the
Modify Options Dialog: Signal Analysis page. The signal analysis options for this data series can be overridden from the Signal Analysis Options Dialog.

·      Entry Signal for First Trade Does Not Appear in Chart View

By default, entry/exit signals that are displayed in the Chart View do not include redundant signals. For example, once an Enter Long signal has been displayed, no Enter Long signals will be displayed until after an Enter Short or Exit Long signal has been displayed. Therefore, if the signal analysis date range does not start at the beginning of the data, a redundant entry signal that is not displayed in the chart may start the first trade.
Redundant entry/exit signals can be displayed in the Chart View by changing the display characteristics for a charted entry/exit signal from the
Change Display Properties Dialog.

Why does the percent equity value for my signal analysis field not match the value I had in TradingSolutions Version 1.0/1.1?

Signal analysis fields that report equity values as a percentage may report different values than those reported in TradingSolutions Version 1.0/1.1. The difference appears specifically on the bar that signals are generated. This is because the entry commission was previously charged at the end of the bar on which the signal was generated. Now, the entry commission is charged at the start of the following bar. This value is otherwise unchanged.