Let's return to my analogy between sports and trading. I live in Toronto, Ontario (that's in Canada for all my foreign friends) and as most people know, hockey is a religion in this part of the world. Our beloved Toronto Maple Leafs are currently in a 40+ year drought in their quest for the holy grail of hockey, the Stanley Cup. To make matters worse, not only have the Leafs not won the ultimate prize in hockey, they have, for the past four years, been unable to even make it into the 1st round of the playoffs. A week ago, after the regular season ended, the NHL (National Hockey League) held their annual lottery draft, where teams, through a random selection, are "positioned" from 1 thru 30, as to who gets the first selection in the upcoming draft for the 2009-2010 season. So, I know you are asking yourself what the similarity is between this event and trading - and in particular weighting. The process is not only similar - it is identical in its approach. Here's how. In the NHL and I believe the same approach is also used in the NBA, and the NFL, teams with the worst regular season records have the greatest chances of getting the 1st pick (or the 2nd, 3rd, 4th and so on) at the draft where all the new young talent is about to enter their professional sport. How do they do this? Simply put, they "weight" the outcomes. We are all familiar with the "throw your name in the hat" approach, well this is exactly what they do in the NHL except they use a large rotating ball or drum, much like the lottery corporations do and in that drum they place balls representing each different team. The outcome is completely random of course, however, teams with worst records have multiple balls in the drum thereby increasing their odds as to getting the top picks. This is exactly how it should be too, otherwise you have teams, regardless of the sport, which dominate the league year after year. This process brings about balance by offering an edge to the weaker teams - giving them a better chance at the top draft picks to strengthen their franchises in the upcoming season.
I have applied the identical methodology into the F-Shift Forecaster ©. I populate the platform with actual returns, again regardless of the time frame, monthly, weekly or daily. Think of a deck of playing cards. There are four cards for each individual card ranging from 2 thru to the Aces. Pulling a King has the same statistical probability as the other cards given there are four of each. Now let's go ahead an open a 2nd deck of cards and take the 4 kings out and add them to our original deck. We have now "weighted" the deck so that the probability of pulling a King from the deck has increased - we now have 8 kings in a deck where every other card only has a maximum of 4. In the F-Shift Forecaster ©, I populate the platform with actual historical data - let's use daily for illustrative purposes. We know that recent data points are far more relevant that older data so in order to compensate for this "importance" we count the most current data multiple times relative to the older "stale" data. The modelling capabilities of this platform really do take this concept to the next level by allowing the end user to determine his or her own weighting metrics. By using the built in spinners you can run an unlimited number of "what-if" scenarios not only by the range of weighting (say for example 20% weighting to the 5 most recent trading days) but you also determine how many weighting "groups" you wish to impose on the data. Allow me to explain. The platform is designed to accept 60 data points - daily, weekly or monthly is irrelevant. If we choose to model 10 (the maximum number) individual weightings that would leave you with 6 data points per "grouping" (again let's assume daily for this example) that you could apply a weighting factor to. (60/10 weightings =6). You could just as easily wish to only model 5 weightings, which would have you analyzing 5 "groupings" of the 12 most recent days and applying you're weighting to that set.
Finally I have included a "BIAS" component to this platform. Within the F-Shift Forecaster ©, you will find a check box and a spinner which basically factors in your opinion or bias mathematically. If you think a stock has rallied quite hard as of late and you are of the opinion that a selloff or profit taking pullback is at hand, then feel free to model that opinion into your equation prior to running the Monte Carlo simulation. Now each randomly ordered outcome (based on historical data downloaded from Yahoo finance) will have in it, your bias as to what possibly may happen in the next 12 periods. I say 12 periods because that is the way the platform is designed - to forecast where you think the stock, index, futures contract may end up over the next 12 days, weeks or months. Once again this all depends on the time frame you have chosen to analyze.
Between the customizable weighting component and the built in biasing modeler, the F-Shift Forecaster is the ULTIMATE tool for swing and options traders. This platform is a completely new and proprietary way to forecast probable outcomes using actual historical prices, and then re-sampling that same data to gain an edge on your trading portfolio using Excel's "RANDBetween" function otherwise known as a Monte Carlo Simulation. (activated by tapping F9)
Please enjoy the annotated screen shot (click the link below) highlighting the platforms features and capabilities. A series of web tutorials are currently being produced along with a Microsoft PowerPoint tutorial on how to optimally use the F-Shift Forecaster © brought to you by Fulcrum Shift Trading. Future versions are currently being tested with historical implied volatility as well and I will bring you up to date on any new developments.
Fulcrum Shift Trading