PSX Desk
Every evening a computer checks all ~500 PSX shares. Gainer Radar picks the 5 shares most likely to rise +5% or more in the next session, and every pick is graded honestly the next evening. Long-term test portfolios run next to the KSE-100 index.
Prices from Fri 13 Nov 2026
Gainer Radar – next session
i
Each night the AI reads all new PSX company notices and a price table, and picks the shares most likely to rise +5% or more next session. Every reason must point to a real notice or to the price data. Other methods pick too, so we can compare. Paper test only until the scoreboard shows it beats random picks.
Test rule: buy at the open only if the price is not more than +3% above yesterday's close · sell at +5% above the buy price, or at the close · 0.5% costs taken off.
Other methods tonight (for comparison):
i
Hit = the day's high reached +5% above yesterday's close (half of the upper lock). Big hit = +7.5%. Lock = closed at the upper limit. Money per night = average profit of the 5 slots after costs (a slot that could not be bought = 0). A method counts only if it beats Random 5 night by night. About 55 nights (~12 weeks) are needed for a fair verdict; a clearly bad method can be stopped after 20 nights.
Long-term test: can we beat KSE-100?
i
A = 20 KSE-100 shares that rose most in 6 months, are near their 1-year high and move calmly; one third changes each month. B = holds KSE-100 only when it is above its 50-day average; otherwise the money is kept safe, not in shares (earns bank-like profit). All start at 100. Paper test, costs included.
Real money – step by step
Small real money first. More only when the results earn it. Checked every Friday.
i
Test budget = money you can afford to lose. Move up a step only when its test passes. 13-week result 4+ points behind KSE-100 → go down one step. Whole budget down 10% from its best → all to the index fund for 4 weeks. Gainer Radar gets real money only after it beats random picks (~12 weeks).
Step 0 · now
80% index fund · 20% copy of ARadar + swing: paper only
Step 1 · 13 weeks
60 index · 35 A · 5 radarIf A is not 3+ pts behind KSE-100
Step 2 · 26 weeks
40 index · 45 A · 15 radarIf the radar beats random
Step 3 · 52 weeks
Grow the budgetA still ahead after costs + tax
Swing watch list (2–10 day trades)
Learning list, paper only
Market info
Indices, news, locked shares, most traded
Shares at the daily limit (±10%) – at +10% nobody sells, so a buy order does not fill
Most traded today
How it works
4 steps, every night
Save prices. Open, high, low, close and volume of all ~500 shares after the close.
Keep safe shares only. Price Rs 10+, at least Rs 50 million traded per day on average. Thin shares are easy to push and hard to sell.
Pick. AI research (main) + buying pressure + biggest risers + random + old score. Saved before 09:00 with a time proof.
Grade. Next evening: bought at the open? Sold at +5% or the close? Money made after costs, night by night, against Random 5.
Archive – older tests
Rules no longer used, kept for the record
| 5-year test (Jan 2022 – Oct 2026) | Trades | Avg per trade | vs KSE-100 |
|---|
| Old swing score (until 21 Oct) | 3,126 | −0.05% | −0.80% |
| Random 5 shares a day | 5,813 | +0.03% | −0.61% |
| New swing score + KSE-100 filter | 3,653 | +0.22% | −0.51% |
No swing rule beat KSE-100 in the 5-year test. That is why the main focus moved to Gainer Radar and the long-term portfolios.
Not investment advice. Research and learning tool. Results are paper tests unless marked "real". Past results do not promise future results. Every trade can lose money; a share at the upper lock can fall to the lower lock.