Order Online Power Walk Forward
Optimizer
Walk Forward
Performance Explorer
Walk Forward
Metric Explorer
Walk Forward
Input Explorer
Walk Forward
Surface Explorer
Key Daily & Intraday
Trading Strategies
Nth Order Fixed Memory
Polynomial Strategy
Nth Order Fading Memory
Polynomial Strategy
End Point Fast Fourier
Transform Strategy
Goertzel DFT
Strategy
Five Parameter
Parabolic Strategy
Dennis Meyers
Working Papers
Algorithmic Trading Strategies For TradeStation, MultiCharts 64bit & NeuroShell
The Power Walk Forward Optimizer (PWFO) is a cutting-edge automatic walk forward/out-of-sample analysis program that eliminates the ad hoc curve fitted performance and data mining results produced by combinatorial and genetic(grail) optimization of strategy input values on spurious price movements(noise). Included in the in-sample section output for each set of input parameters are 30 new robust and superior performance metrics. The in-sample and out-of-sample periods are user selectable. The PWFO can generate up to 500 different in-sample and out-of-sample date files in one TS run. Statistically speaking, walk forward out-of-sample (oos) analysis must be performed over many(>30) in-sample/oos sections to be statistically valid.

Click here to for a User's Comparison Between the TS Walk Forward Optimizer(Grail) and the PWFO

Videos On How To Use The PWFO 1. How to setup and run the PWFO  2. The PWFO output files
Available for TradeStation, MultiCharts and NeuroShell Pro
eqwfmeThe Walk Forward Metric Explorer (WFME) reads all files generated by the PWFO and searches each PWFO file for performance Metrics that generate statistically best average out-of-sample performance. The Top N Metric filter chooses the PWFO file rows that contain then Top N (N=10 or 29 or etc) values of a PWFO Performance Metric. From the N rows chosen, the WFME chooses the maximum of another PWFO performance metric. This allows two performance metrics in the in-sample section to be used to find the strategy inputs that give the best out-of-sample returns. Experience has shown that the use of only one in-sample performance metric to choose strategy inputs from the in-sample section does not produce good out-of-sample results. The Top N criteria ranges are user selectable generating many filter searches in one run. The WFME is a stand alone exe program that is super fast and automatically displays it's extensive statistical results, equity plots and strategy inputs from each filter in Excel. In addition, using modern "Bootstrap" techniques, the WFME calculates the probability of whether or not each filter's out-of-sample results were due to chance.
Video: How To Use The Walk Forward, Out-Of-Sample PWFO Metric Explorer
eqwfpeThe Walk Forward/Out-Of-Sample Performance Explorer (WFPE) reads all files generated by the PWFO and searches the in-sample performance metrics in those files for those performance metrics that generate the statistically best average out-of-sample returns. The WFPE eliminates all cases in the PWFO test files that do not meet certain Profit Factor (PF), Losers in a row (LR),and PWFO Performance metrics criteria . The PF, LR and METRIC criteria/search ranges are user selectable. The WFPE is a stand alone exe program that is super fast and automatically displays it's extensive statistical results, equity plots and strategy inputs from each filter in Excel. In addition, using modern "Bootstrap" techniques, the WFPE calculates the probability of whether or not each filter's out-of-sample results were due to chance.
Video: How To Use The Walk Forward, Out-Of-Sample Performance Explorer
eqwfpeThe Walk Forward/Out-Of-Sample Input Parameter Explorer (WFINP) reads all files generated by the PWFO and searches the performance metrics of profit factor(PF) and losing trades-in-a-row(LR) in those files for those performance metrics that generate the statistically best average out-of-sample returns. The WFINP eliminates the curve fitted results by filtering out from the PWFO file's in-sample sections those strategy inputs that have Profit Factors(PF) greater than x and that have losing trades in a row(LR) of greater than y. The WFINP then determines which strategy filtered inputs generate the statistically best average out-of-sample returns. The PF and LR criteria are user selectable so you can choose which PF and LR values suit you. The WFINP is a stand alone exe program that is super fast and automatically displays it's statistical summary and results in Excel. In addition, using modern "Bootstrap" techniques, the WFINP calculates the probability of whether or not each filter's out-of-sample results were due to chance.
3DSurfThe Walk Forward/Out-Of-Sample Metric Surface Explorer(WFSE) reads each of the files generated by the PWFO and calculates the flattest metric surface plateaus vs the input parameters for each performance metric's (Total Net Profits, Profit Factor, etc) surface. The WFSE can find the flattest surface for up to a 7 dimensional(7D) surface. A 7D surface consists of six input parameter axis vs a performance metric variable. A performance metric flat surface plateaus represent robust input parameters that have a greater chance of producing good out-of-sample results. The WFSE then finds the out-of-sample profits associated with each in-sample section surface's flattest plateau (minimum gradient) for each file. The WFSE sums each file's out-of-sample results for each of the surface's minimum gradients and finds the statistically best out-of-sample returns for each performance metric surface. The WFSE is a stand alone exe program that is super fast and automatically displays it's extensive surface statistical summary and results in Excel.
Using fast advanced mathematical rocket science algorithms, the price series is modeled using an nth order fading memory polynomial of the form:  price(t) = a0(t)+a1(t)*t+a2(t)*t2+a3(t)*t3+a4(t)*t4+...+an(t)*tn  The an(t) coefficients are updated recursively with each new price bar and then used to give the polynomial's next bar forecast of price,velocity and acceleration. As our working papers demonstrate the nth Order Fading Memory Adaptive Polynomial is an effective strategy for trading stocks, futures and currencies. Available for TradeStation, MultiCharts and NeuroShell Pro
Using fast advanced mathematical rocket science algorithms that use discrete orthogonal polynomials, the price series is modeled using an nth order polynomial of the form:   price(t) = a0+a1*t+a2*t2+a3*t3+a4*t4+...+an*tn  The an coefficients are recalculated at each new price bar and are then used to give the polynomial's next bar forecast of price,velocity and acceleration. As our working papers demonstrate the nth Order Fixed Memory Adaptive Polynomial is an effective strategy for trading stocks, futures and currencies. Available for TradeStation, MultiCharts and NeuroShell Pro
The Adaptive n Cycle Goertzel Discrete Fourier Transform (nCycleGZ) super fast DLL finds the N(user selectable) cycles(frequencies) with the highest amplitudes at each price bar, and creates a x bars ahead (x is user selectable) noise filtered projected momentum curve. This process gives a more robust noise filtered signal than a single frequency (dominant cycle) procedure. You are no longer constrained to using only a single frequency. The nCycleGZ algorithm is faster and superior to MESA in finding cycles in noisy price series. As our working papers demonstrate the nCycleGZ is an effective strategy for trading stocks, futures and currencies. Available for TradeStation, MultiCharts and NeuroShell Pro
The EPFFT super fast DLL takes the FFT at each price bar, filters the noisy price series using a unique noise filter in the frequency domain, and creates a one bar ahead noise filtered projected price. The EPFFT DLL produces an adaptive broadband (many frequencies) noise filtered signal. This process gives a more robust noise filtered signal than a single frequency (dominant cycle) procedure. As our working papers demonstrate the EPFFT is an effective strategy for trading stocks, futures and currencies. Available for TradeStation, MultiCharts
The five parameter parabolic adds a noise filter and changeable starting stop value that minimizes the whipsaw losses that can occur with the regular parabolic indicator. Here this new system is applied to stock and Futures prices to minimize the noise process. Available for TradeStation, MultiCharts


top Top | Home