Run simulations, compare the perfect path to model predictions with a pinch of common sense, and see how your strategy would have performed before risking capital.
Why backtesting matters
Dividend capture can be highly profitable, but only when you do it consistently and playing with the upper hand. Backtesting lets you test ideas against historical reality, so you know whether a strategy is promising or just wishful thinking. Instead of guessing, you can measure.
Two simulation types we offer
We provide two complementary simulation modes:
- System generated simulations — default parameters run automatically for each week, month, and year.
- User custom simulations — you choose thresholds and rules to test your own hypotheses.
Both let you compare the theoretical best outcome against what our prediction plus some filters and upfront logic would have produced over the same period. No hidden magic, you will see in detail exactly what we do
The Perfect Path explained
The perfect path is the theoretical maximum-earnings route. It evaluates every possible dividend combination under two simple constraints: at most one trade per day, and all capital allocated to a single event each time. The simulation then finds the sequence of events that yields the highest cumulative return for the period.
This is a benchmark. It shows what perfect selection would have returned, and it defines the ceiling against which prediction performance is measured.
The Prediction Path explained - Candidate selection
The prediction path simulates how our model would have traded. It only considers events where the model predicted same‑day recovery above a configurable probability threshold (default 50%).
Each day the simulation filters candidates by yield and probability: first we keep dividends above a yield threshold tied to the day’s highest yield, then we select the single event with the highest model probability. Simple and transparent.
Example with filters: 85% probability and 90% yield thresholds. Highest yield dividend of the day is 4%, this means the candidate pool will be limited to those with yield above 90% of that 4%, which is 3.6%
| Yield | Probability | Candidate | Winner |
|---|---|---|---|
| 0.85% | 92% | No | No |
| 4.00% | 87% | Yes | No |
| 0.40% | 99% | No | No |
| 3.75% | 95% | Yes | Yes |
| 2.60% | 81% | No | No |
How wins and losses are simulated
If a prediction succeeds, we assume the trade is closed at the pre‑ex price, meaning the buy/sell operation nets no extra profit or loss and the dividend is the realized gain. If a prediction fails, we simulate selling the next day at the median of the day’s OHLC prices and compute the net result as dividend minus buy/sell loss.
These rules keep the simulation conservative and easy to interpret: successful predictions capture the dividend cleanly; failed predictions reflect realistic trading friction and price movement. Those cases can return a negative number, a real loss.
In any case, the initial capital will always be the same through the entire simulation, we don't reinvest the earnings on the following operations
Interpreting results and the capture rate
Once the two paths are computed, we compare outcomes. For example, if the happy path returned 22% and the prediction path returned 12%, the capture rate is 12 × 100 / 22 = 54%. That number tells you how close the model and the logic applied came to the theoretical maximum for that period.
We aggregate simulations by week, month, and year so you can spot trends. Explore the full analysis: The Perfect Hunting: Finding Your Optimal Threshold & Filters.
Practical use and next steps
Backtesting is a tool, not a promise. Use system simulations to get a baseline and run custom simulations to test your thresholds and risk rules. There you can also narrow the candidates to specific Exchanges or to your own Watchlist. Adjust the probability cutoff, change yield filters, and see how the capture rate and returns move.
We call the historical backtest The Perfect Hunting. For forward‑looking scenarios we offer a live simulation called Stalk & Strike, able to create a Prediction Path for the next week. Both simulations apply the very same candidate selection logic. Either way, test before you trade: it’s the fastest way to turn intuition into evidence.
For detailed guidance on setting up your own custom filters and parameters, see Build Your Own Backtest: Custom Simulations Guide.
