@mychocoalbab: Rising Star ##risingstar #aasia #chocoalbab #trending #tiktokmalaysia

Choco Albab MY
Choco Albab MY
Open In TikTok:
Region: MY
Saturday 01 August 2026 03:34:32 GMT
4687
200
6
21

Music

Download

Comments

ciklady9
Nekma Choco :
🥰🥰🥰wowwww
2026-08-01 09:59:19
0
sangmentari955
Sang Mentari955 :
terrrbaaiiiikkk choco Albab 🔥🔥🔥🥰🥰🥰
2026-08-02 04:25:02
0
rahmah.ahmad61
RAHMAH AHMAD :
2026-08-02 09:43:35
0
hasmawati1907
Asma Racun Online :
wao
2026-08-02 05:14:07
0
aqil.khai4
Mom's 4 NurMuhammad :
Choca albab ni syarikat cina org mamak ke?
2026-08-02 11:23:16
0
aziyah_hussin
Azie' shop :
👍👍👍
2026-08-10 08:37:34
0
To see more videos from user @mychocoalbab, please go to the Tikwm homepage.

Other Videos

Walk-forward backtest optimization tested on 16 years of E-mini S&P 500 futures data. We optimized three strategies on a rolling 2-year window, then traded the chosen settings blind on the next 6 months, and rolled that forward 28 times each. Donchian breakout, moving average crossover and Connors RSI2. 138 parameter combinations, re-selected at every step. In training the selected settings made $1,458,122. Out of sample the same settings lost $604,740. That is a pooled out-of-sample to in-sample ratio of -0.41, in all three strategy families. The edge did not shrink, it inverted. The part we did not expect: for every fold we also scored a parameter set picked at random from the same grid. Across 84 paired tests the optimized choice lost to the random one by about $5,067 a fold (paired t-test p = 0.027), and lost to the grid median by $5,486 (p = 0.0003). Optimizing did worse than not bothering. Costs were measured, not assumed. 73,737,856 real top-of-book quotes across 16 days spread over 2010 to 2025 give a 0.2606 point average spread and a $13.03 round turn. At double that cost the ratio is -0.44. 7 of 15 years were positive, so this is not one bad year, and buy and hold made $310,148 over the same windows and beat all of it. Validation label: walk-forward out of sample, no evidence of exploitable edge. The 5.6 million minute bars and the 73.7 million quotes behind this are included with QuantPad. No vendor account, no CSV hunting. #QuantPad #backtesting #algotrading #quanttrading #futurestrading
Walk-forward backtest optimization tested on 16 years of E-mini S&P 500 futures data. We optimized three strategies on a rolling 2-year window, then traded the chosen settings blind on the next 6 months, and rolled that forward 28 times each. Donchian breakout, moving average crossover and Connors RSI2. 138 parameter combinations, re-selected at every step. In training the selected settings made $1,458,122. Out of sample the same settings lost $604,740. That is a pooled out-of-sample to in-sample ratio of -0.41, in all three strategy families. The edge did not shrink, it inverted. The part we did not expect: for every fold we also scored a parameter set picked at random from the same grid. Across 84 paired tests the optimized choice lost to the random one by about $5,067 a fold (paired t-test p = 0.027), and lost to the grid median by $5,486 (p = 0.0003). Optimizing did worse than not bothering. Costs were measured, not assumed. 73,737,856 real top-of-book quotes across 16 days spread over 2010 to 2025 give a 0.2606 point average spread and a $13.03 round turn. At double that cost the ratio is -0.44. 7 of 15 years were positive, so this is not one bad year, and buy and hold made $310,148 over the same windows and beat all of it. Validation label: walk-forward out of sample, no evidence of exploitable edge. The 5.6 million minute bars and the 73.7 million quotes behind this are included with QuantPad. No vendor account, no CSV hunting. #QuantPad #backtesting #algotrading #quanttrading #futurestrading

About