Awaiting data
Every assumption has a concrete quarterly target.
When earnings arrive, we check: did the target get met?
If the AI said “on track” and the target was met — correct.
If the AI said “in trouble” and the target was missed — also correct.
Mismatches are failures. Higher accuracy = the AI is reading the news correctly.
After first earnings cycle (~Jun 2026)
-8.90%
long-short spread
Hit rate (Buy):
21.1% (4/19)
Hit rate (Sell):
35.7% (5/14)
Buy return:
-6.86%
Sell return:
+2.04%
⚠ Unstable signals: HDFCBANK (6.2/90d), RELIANCE (6.2/90d), INDIGO (5.3/90d), MARUTI (10.7/90d), LT (4.5/90d), SBIN (5.3/90d), ITC (7.1/90d), GOOGL (4.4/90d), AAPL (6.2/90d), MSFT (5.3/90d), AXISBANK (10.7/90d), TATASTEEL (7.1/90d), TRENT (4.5/90d)
Spread > 0 and hit rate > 60% = success.
One vote per stock per bucket. Min 7-day holding.
After 3-6 months (~Sep 2026)
+4.49%
blended alpha (gross)
+2.88%
blended alpha (net of costs)
Portfolio:
-5.76%
Benchmark:
-8.67%
Alpha (gross):
+2.91%
Turnover:
+962.01%
Cost drag:
-1.92%
Alpha (net):
+0.98%
Portfolio:
+25.82%
Benchmark:
+13.42%
Alpha (gross):
+12.41%
Turnover:
+98.21%
Cost drag:
-0.05%
Alpha (net):
+12.36%
Success = the AI portfolio beats at least one market index.
Uses the actual fund NAV — Buy-rated stocks receive
higher conviction-weighted allocation, Hold/Sell stocks
receive the remainder equally. Rebalanced daily.
Net alpha accounts for estimated transaction costs
(India: ~0.20%/round-trip, US: ~0.05%/round-trip).