I run a few versions of Monte-Carlo Realty Check on my strategy, sharing the results with you.
You can find an excellent blog on this issue here:
Monte Carlo Reality Check As jcl wrote in the above post:
Therefore, neither MRC nor WRC can give absolute guarantee that a system works when it passes the check. But when it does not pass, you’re advised to better not trade it with real money.
I would appreciate to hear from you, insights and your experience on this regard.
My strategy is a rotation one, running once a week, long only, no leverage.
I optimized two global variables: lookback period and number of assets to trade.
Those are the results:
1. MRC with SUFFLE - Test only, no train, uses default parameters
-------------------------------------------
P-Value 2.0%
Result is significant
-------------------------------------------
Win 75814$ MI 2111$ DD 12690$ Capital 100934$
Trades 65 Win 50.8% Avg +479.0p Bars 33
CAGR 20.75% PF 2.94 SR 1.75 UI 7% R2 0.88
Time 00:03:39
See the attached histogram plot
2. MRC with BOOTSTRAP - Test only, no train, uses default parameters
-------------------------------------------
P-Value 1.9%
Result is significant
-------------------------------------------
Win 76200$ MI 2122$ DD 5109$ Capital 85135$
Trades 76 Win 59.2% Avg +598.6p Bars 28
CAGR 20.84% PF 2.87 SR 1.47 UI 13% R2 0.70
Time 00:03:47
3. MRC with SUFFLE, 2 parameters trained
Run MRC..........................
Read MRC_1.par
Read MRC_2.par MRC_3.par MRC_4.par MRC_5.par MRC_6.par MRC_7.par MRC_8.par MRC_9.par MRC_10.par MRC_11.par MRC_12.par MRC_13.par MRC_14.par
-------------------------------------------
P-Value 2.0%
Result is significant
-------------------------------------------
Win 82974$ MI 2311$ DD 5544$ Capital 61839$
Trades 68 Win 58.8% Avg +571.6p Bars 29
CAGR 22.37% PF 3.88 SR 2.03 UI 4% R2 0.94
Time 06:22:37
4. MRC BOOTSTRAP, 2 trained parameters
Run MRC..........................
Read MRC_1.par
Read MRC_2.par MRC_3.par MRC_4.par MRC_5.par MRC_6.par MRC_7.par MRC_8.par MRC_9.par MRC_10.par MRC_11.par MRC_12.par MRC_13.par MRC_14.par
-------------------------------------------
P-Value 1.9%
Result is significant
-------------------------------------------
Win 15548$ MI 433$ DD 5863$ Capital 77991$
Trades 72 Win 43.1% Avg +197.1p Bars 28
CAGR 4.95% PF 1.45 SR 0.43 UI 40% R2 0.00
Time 05:06:57
The p value for all versions is around 2%, which is very good.
The CAGR for the 3 version are quite the same, for the last version it drops a lot, I don't understand what that means.
David
P.S. I don't have Zorro-S license, so the run time is very long
