What the Platform Backtest Actually Checks: Signals, Bias and Baseline
The backtest section is easy to misread: it looks like a school report on whether astrology is accurate, when it actually does something narrower. It checks whether the events you recorded tend to co-occur with certain chart signals in the past. Done carefully, that tells you something about relative strength between signals; overread, it becomes retrospective pattern-fitting.
What a test actually checks
We take your natal chart as the starting point and select its computable timing signals: the theme of the current Firdaria time lord, the sign and aspects of the progressed Moon, and the transiting aspects that land on your natal sensitive points. For each period of life you entered, the check is whether these signals cluster within a few days around the events you marked. What is really being tested is co-occurrence, not whether the event would have happened anyway. Windows with dense, clear aspects do overlap events more often; but whether that overlap has any causal direction is something the test does not answer.
Where the bias comes from
Backtesting trips over two biases. Selection bias: the events you enter are the ones you remember, the significant and emotional and recent ones, while ordinary days quietly drop out of the sample, which therefore leans toward the very windows where something was happening. Hindsight bias: the mind finds it almost impossible to settle on a conclusion first and look for evidence later, and with many aspects and wide orbs almost any event can be matched after the fact. The platform fixes thresholds and date windows to reduce both, but it cannot remove the simple fact that you already know the outcome when you enter the data.
Recurrence within the sample is not prediction
A signal that hits seven or eight out of ten recorded events may sound impressive, and yet it answers only one question: does it recur within the events you selected? It leaves the more basic question open: how often does the same signal appear during the ordinary days when nothing was happening? If the background baseline is already dense, a high hit rate just reflects a crowded signal. Real prediction would require showing that the event rate rises above baseline when the signal appears. The backtest section deliberately does not do that; it reports relative strength and labels itself a retrospective observation.
How to actually read the section
Read it as a rough ranking of signals, not as a verdict. When several timing systems point at the same period, tradition already tends to weight the time lord first; if the backtest also shows stable co-occurrence for it, that gives you a reason to prefer it among competing candidates. Conversely, a lukewarm rate for one signal does not mean it is useless: your sample may simply be too short. The best habit is to treat the result as one more hint, lay it back beside transits, progressions and house readings, and let your actual situation and choices stay the deciding voice.
In short, the backtest tells you which signals look salient within your own sample of life. It is neither a scientific experiment nor a prediction tool. Treat it as a sieve that suggests what to look at first, then decide with common sense, evidence and professional judgment. Seen within its own limits, it may be the most honest instrument the platform offers.
FAQ
Does a positive backtest mean astrology works?
A two-part answer. The backtest records co-occurrence, and co-occurrence can be amplified by selection and hindsight bias, so a positive result is not scientific proof on its own; it only shows that some signals were more salient among the events you entered. The platform treats it as an observational aid and never presents the result as proof.
Why does a signal work sometimes and not others?
Because no signal dominates at every moment. The backtest reports a tendency within your sample; individual cases carry noise, and events are shaped by countless real-world variables, with chart signals being only one interpretive thread. Reading a single miss as a failure means misusing a statistical tool as an oracle.
Can I use the backtest to plan important dates?
We would advise against it. The backtest looks backward; it has no forward-looking predictive power. For important plans, rely first on real-world evidence, professional advice and your own judgment, and treat astrological reading as a complementary reflective view.