NEPSE Signal Track Record
Daily scan signals
| Signal | Signals | After 10 days | vs NEPSE (10 days) | Rose (10 days) |
|---|---|---|---|---|
| Strong Buy | 21 | — | — | — |
| Buy | 36 | — | — | — |
| Hold | 51 | — | — | — |
| Sell | 30 | — | — | — |
| Strong Sell | 13 | — | — | — |
Backtest and calibration
Tuned on 2025-12-18 to 2026-06-11, checked on 2026-07-10 to 2026-10-02.
The calibrated rules did not do clearly better on the days they weren’t tuned on, so the daily scan keeps its current rules.
How it works
After every daily scan, each share’s signal is saved with that day’s closing price. 5, 10 and 20 trading days later the site records how far the share moved, and how far NEPSE moved over the same days.
The backtest replays the engines on every trading day of up to five years of stored prices, using only what was known that evening. The calibration then tries different weights for each part of the technical score (RSI, MACD, Bollinger Bands, trend, volume) and for the RRG score, and different Buy levels, on the older 70% of the days. The best rules are then checked on the newest 30%, which the tuning never saw.
New rules are used only if they beat the current ones on those newer days by a clear margin (at least 0.1 percentage points over 10 days, and steadier). Weights stay between 0 and 2, so a part can count less or more, but never the opposite way.
Past results, real or replayed, do not predict future returns. Backtests leave out trading costs, price slippage and liquidity limits. This is education, not investment advice.