The testing is accurate based on the methodology used. The selection process at each rebalancing is the same as that used in the Screener, using point-in-time data for the date of rebalancing. The entry and exit price for each instrument is the closing price for the day of the rebalancing; the return for the period determined by the difference between the total worth of all investments at the beginning and end of each period. The dates of rebalancing depend on the date of the test. The backtest starts 5 years (260 weeks) prior to the most recent Friday at the time of the test, and moves forward in 13 week increments, rebalancing at the end of each period.As such, identical tests run on different weeks will see different results, as the rebalance date will fall on a different date and different instruments may be selected, having different entry and exit values. In addition to the timing, the 'accuracy' of the results, in terms of matching expectations, is also highly dependent on the criteria used. A test run one week with poor returns may show excellent returns when run at a different time. A small handful of instruments selected over the course of a backtest that performs exceptionally well (or poorly) will shift the overall results in the direction of their movement. This could result in a screen designed to find bullish instruments turning up more bearish ones instead. Our backtesting does filter out outliers in order to forcus on more likely returns. Instruments do not always move in predictable ways. The methodology of the backtests is followed rigorously. The results need to be looked at within the context of the methodology and the realization that a strategy does not always work as expected.

The methodology for the backtesting is a 3-month buy and hold strategy.The test is run for 5 years, ending on the most recent Friday.So for example, a test run on April 27, 2011 will run for the 5 year window starting Friday April 28, 2006 to Friday April 22, 2011.It will run its first screen for Friday, April 28, 2006, hold the matching instruments for 13 weeks, rebalance on Friday, July 28, 2006, hold those for 13 weeks, rebalance on October 27, 2006, etc. Here is a breakdown of the process: At the beginning of the test period, the top 10 matching instruments are selected (or as many as match, if the number is less than 10). The beginning of the test window being 5 years prior on the Friday. A position is taken in each instrument, at a uniform distribution of available funds. Since all returns are measured in terms of percentage return, the total value of this initial pool is arbitrary. The instruments are held for 3 months (13 weeks), with the performance of the pool measured on weekly intervals for reporting purposes. At the end of 3 months, on the Friday, the instruments are deemed sold at the closing price. The process is repeated with a new set of instruments, distributing the total worth of the instruments from the previous period across the newly selected instruments (and so on), for the duration of the 5 year test period. There are no stops or any other adjustments to the instruments selected between rebalancing periods.

The criteria sliders allow for easy selection of criteria value ranges. This way you don't need to guess what values might represent strength or weakness in a stock, you can simply adjust the slider to choose the highest, lowest, or whatever other range makes sense for your strategy. Depending on the criteria, the minimum and maximum values are displayed as a number or percentage. After isolating the outliers, the remaining values are evenly distributed across the slider. If you want to look for stocks with values outside the main slider body, there are three ways to do so: enter the desired limit in the min or max entry field; enter 'min' or 'max' in the respective field to remove the limit entirely; or drag the appropriate slider control to the extremity. The 0 through 5 values for Technical Events signify the 'strength' of the event, where a higher value represents a stronger event. This is based on a Recognia developed algorithm that measures the initial strength of an event (based on metrics for that event) and the point in time the event occurred. The strength of the event reduces over time, meaning that an event is most relevant at the time it occurs, but becomes less relevant as time passes. Since the value for an event is a based on two factors - initial strength and time since the event occurred - a value can be arrived at in a number of ways. For example, an event with a value of 3 might have originally been a strong (perhaps starting with a strength of 5) but after a long enough period of time passed the strength was downgraded to 3.Or perhaps that same event was not as strong initially so its starting value was 3. The slider for Technical Events is there for a user to filter the most recent and strongest events (via high slider values) or all active events (by selecting the full range of values).

Alerts will notify you as the top 10 list of stocks changes over time. Strategy Builder isn't just a stock screening tool; it's designed to help you follow a well suited strategy over time. When you turn on alerts for a screen, you will receive emails notifying you about which stocks have dropped off the top 10 list, and which stocks have newly made the cut. This saves you from constantly returning to find out the latest matches. When you decide what you want, it will just be emailed to you! To turn on alerts, you have to save the alert as your own so that it will appear in the "My Saved Screens" list. If you are building a new screen or looking at someone else's, you can save it by clicking the "Save Screen" button that appears below the list of criteria included in the screen. This will open a dialog box where you will see a checkbox for "Email me any changes to the 'top 10' matches". Click the checkbox to activate the alerts. You can decide if you want your updates Daily, Weekly, or Monthly depending on how often you revisit your investments. If you've already saved a screen and want to turn on alerts, just open the screen by clicking on its name in the "My Saved Screens" box. Scroll far down on the page, below the list of criteria, and find a link called "Edit Information". This will bring you the dialog described above, and you can check the box for "Email me changes to the 'top 10' matches".

Backtesting is performed on a 5 year window on the latest available data, which means the 5 year window changes from week to week. So it's quite likely there will be some change in results from week to week. All backtests are run on weekly intervals, so they run from the most recent Friday, and then go back 5 years.The screen that is being tested is run once on a date 5 years ago, the instruments matching the screen are bought and held for 13 weeks, sold, then the screen is run again and the new instruments matching the screen are bought -- and so on, for the 5 year period. As the weeks pass, the starting date of the screening and backtesting changes and this, combined with the buy and hold period of 13 weeks, means that the historical screens never get run on the exact same days. It is quite possible that the latest backtest holds a completely different set of instruments ... and, even if they were the same, they would be held for different periods in their history. Considering this, it is not only possible to get different results from week to week, it's common. We believe running on the latest data is important. So to focus on a more likely returns figure, our backtesting filters out the outliers. Stocks that have an annualized return of over 200% over the test period of 13 weeks are rejected. Extreme results may still occur, but they aren't created by single extreme instruments. Instead they are created from a group of instruments with big returns, but not quite as extreme.

Strategies can have vastly different returns depending on the timeframe being evaluated. For example, while writing this response I am looking at a strategy called High Return on Equity. The Annualized Return is -7.7% and the long-term 5 Year Return (based on buying and holding over 5 years with cumulative effects) is -33.2%. However if you look at the near term analysis: 3 year return is +46.7%, 1 Year Return is +0.9% and 6 Month Return is +12%. The strategy suffered during the bearish market periods and performed respectably as the market recovered, and in the last 6 months the returns outpaced the selected benchmark index (in this case DJ Industrial Average). There isn't a perfect strategy for investing that will always provide the best returns. A prudent investor will continuously monitor the effectiveness of the strategies they're using and adapt them as market conditions change. For example, some strategies work best during trending periods, and not so well during sideways markets. Some work best during periods of strong economic growth, and not so well during recessions. We might not always recognize the shifts in market conditions that affect the usefulness of a strategy, but Strategy Builder can be used to tell us how the strategy has been working lately versus other time frames. Returns data is updated on a weekly basis to incorporate the results from the latest week, so you can see that the returns data changes over time. Sometimes these changes are minor; sometimes they can move a strategy from a positive to a negative returns situation.If a strategy was initially featured because it was positive, and then it becomes negative, then we don't necessarily remove it. It might become interesting again in the future.