Add new comment

Screen improvements with diversification

I recently backtested all the standard screens from 1997-08-31 using a 28-day/monthly trading cycle both with and without the new MAXEQUALfunction recently introduced in the backtester (holding top 10,holding cash whenever there were not enough picks).

Most screens (with a couple of exceptions) seems to benefit - as measured by the CAGR/GSD ratio and/or UPR ratio (using a MAR of 1%/month).

I ran the screen first as normal, and then by adding MAXEQUAL(1,[SI Sector]) after the final sorting step in each screen.

','
I recently backtested all the standard screens from 1997-08-31 using a 28-day/monthly trading cycle both with and without the new MAXEQUAL function recently introduced in the backtester (holding top 10, holding cash whenever there were not enough picks).

Most screens (with a couple of exceptions) seems to benefit - as measured by the CAGR/GSD ratio and/or UPR ratio (using a MAR of 1%/month).

I ran the screen first as normal, and then by adding MAXEQUAL(1,[SI Sector]) after the final sorting step in each screen.

Comments below whenever it looks like something worth commenting:

78RPM, Eastwood, Quality_Earnings - typical momentum screens. Seems to benefit significantly. Exception is FCF-26.
CANSLIM-26 - sorts on momentum, but maybe not a typical momentum screen. Still the diversified version looks a lot better.
Commodity - Diversification for a non-diversified screen works bad. It only held 3 stocks on average.
Fried-500 - Value screen. Slight improvement.
Gentle Screamers family. Slight improvement (if you want GSD down instead of chasing a few percentage points of CAGR).
High_Relative_Value. Value screen. Does not seem to benefit from diversification. Sector momentum possibly a factor?
Generally, CAGR _and_ GSD goes down for most screens.

To see if I could find any confirmation, I ran a second test using all possible starting weeks from 2003-01-03. That means that roughly 75% of the samples was not tested in the first run (one of the four weeks in the all possible starting weeks for the monthly cycle was tested in the initial run from 1997-08-31);

78RPM, Eastwood, Quality_Earnings. I would call it a confirmation.
CANSLIM-26. Definitively confirmation.
Fried-500. Improvement is significantly better on the all starting weeks run.
Gentle Screamers family. Still slight improvement. Confirmation.
High_Relative_Value. Same as before. Confirmation of no improvement.
From the all starting weeks run it seems that most screens improve somewhat (measured by looking at CAGR/GSD and/or UPR) with added diversification.

Single run from 1997-08-31, monthly/28-day cycle, one starting week only:

1997-08-31, ytd, top 10, freq 28, cashfornopicks MAXEQUAL (1, [SI Sector])
CA/GS upr CAGR GSD U/D T.o. Cycl CA/GS upr CAGR GSD U/D T.o. Cycl
78RPM 1.10 0.82 34 31 67 71 105 1.41 0.92 38 27 67 74 105
ARS 0.79 0.69 15 19 58 70 105 0.73 0.59 11 15 53 78 105
CANSLIM-26 0.68 0.70 19 28 62 64 105 1.00 0.67 16 16 63 66 105
Commodity 0.67 0.83 18 27 51 58 105 0.57 0.38 4 7 54 67 105
Eastwood 1.74 1.12 33 19 66 78 105 2.05 1.22 39 19 65 79 105
FCF-26 0.72 0.72 18 25 58 48 105 0.41 0.50 9 22 64 46 105
Fried_500 0.68 0.66 17 25 59 17 105 1.00 0.71 22 22 68 23 105
Gentle_Screamers 1.54 1.06 40 26 63 93 105 1.50 0.95 30 20 63 92 105
GS_Mungo 2.00 1.12 40 20 70 90 105 1.78 0.89 32 18 70 90 105
GS_Mungo_LowVoom 1.90 1.18 40 21 67 90 105 1.95 0.99 37 19 71 91 105
GS_Mungo_Voom 1.90 1.11 40 21 70 90 105 1.83 0.91 33 18 70 91 105
High_Relative_Va 2.68 1.49 51 19 77 91 105 1.56 0.95 28 18 68 91 105
Melange 1.04 0.93 47 45 68 58 105 0.82 0.75 23 28 59 65 105
Microcap_Momentu 1.11 1.04 52 47 61 90 105 0.45 0.66 13 29 50 91 105
Net-Nets_Grahami 1.00 1.15 60 60 57 46 105 0.91 1.05 40 44 60 51 105
OptiMan 0.71 0.53 12 17 68 33 105 0.61 0.48 11 18 68 40 105
PIH_Naked 0.38 0.59 18 48 60 97 105 0.43 0.58 16 37 62 95 105
POG 1.29 0.96 54 42 64 60 105 1.22 0.90 39 32 65 65 105
POI 1.14 0.82 33 29 65 63 105 1.22 0.93 33 27 62 68 105
P/S_I_Love_You 1.81 1.09 47 26 70 64 105 1.43 0.87 33 23 70 70 105
Quality_Earnings 1.27 0.81 33 26 66 60 105 1.86 1.03 41 22 71 63 105
S&P_Peg 0.85 0.70 23 27 64 41 105 0.71 0.63 17 24 59 51 105
Shrinkage 1.00 0.74 21 21 68 20 105 0.95 0.65 20 21 70 26 105
Silver_Parachute 0.78 0.75 18 23 60 46 105 0.83 0.67 15 18 61 52 105
Small_Value 1.03 0.87 30 29 61 46 105 1.21 0.83 29 24 68 49 105
Steady_Growth 0.62 0.66 15 24 60 30 105 0.86 0.53 12 14 58 34 105
Turnarounds 0.50 0.62 16 32 61 90 105 0.72 0.63 21 29 67 86 105
Up_5% 1.36 1.00 45 33 67 90 105 1.27 0.86 33 26 67 94 105
Value_at_the_Top 1.15 0.75 23 20 67 70 105 1.11 0.71 21 19 65 78 105
Zweig-26 1.11 0.82 21 19 83 46 105 0.93 0.62 13 14 85 55 105
Current_Ratio_s 0.26 0.66 18 70 51 58 105 -0.04 0.51 -2 55 52 62 105
Dilution_s 0.32 0.68 17 53 52 47 105 -0.21 0.42 -7 33 45 52 105
Negative_FCF_s 0.21 0.63 11 53 56 84 105 0.07 0.55 3 40 50 87 105
The_Mirror_s -0.12 0.53 -6 48 46 88 105 -0.42 0.35 -15 36 46 88 105

Average of all possible startweeks from 2003-01-03:
without MAXEQUAL with MAXEQUAL
CA/GS upr CAGR GSD U/D T.o. Cycl CA/GS upr CAGR GSD U/D T.o. Cycl
78RPM 2.34 1.50 68 29 72 68 41 2.84 1.74 71 25 77 75 41
ARS 1.62 0.93 26 16 67 70 41 1.85 1.02 24 13 63 77 41
CANSLIM-26 1.00 0.73 18 18 66 59 41 1.64 0.78 18 11 72 62 41
Commodity 2.43 1.36 51 21 75 57 41 1.29 0.50 9 7 61 65 41
Eastwood 2.15 1.36 43 20 70 76 41 2.82 1.50 48 17 77 78 41
FCF-26 1.12 0.74 19 17 66 49 41 1.38 0.74 22 16 73 44 41
Fried_500 1.39 0.93 25 18 65 17 41 1.93 1.09 27 14 70 22 41
Gentle_Screamers 2.00 1.22 54 27 66 94 41 2.17 1.26 50 23 74 93 41
GS_Mungo 2.53 1.39 38 15 77 92 41 2.71 1.52 38 14 76 91 41
GS_Mungo_LowVoom 2.31 1.43 37 16 72 92 41 2.64 1.42 37 14 72 92 41
GS_Mungo_Voom 2.25 1.25 36 16 74 92 41 2.92 1.50 38 13 76 92 41
High_Relative_Va 3.44 1.85 55 16 79 93 41 2.82 1.53 48 17 75 92 41
Melange 1.90 1.23 55 29 70 54 41 1.42 0.99 37 26 64 64 41
Microcap_Momentu 1.61 1.34 58 36 62 91 41 1.52 1.22 32 21 64 92 41
Net-Nets_Grahami 1.29 1.11 49 38 66 43 41 1.29 1.12 36 28 68 52 41
OptiMan 0.92 0.53 12 13 65 33 41 1.50 0.80 21 14 69 39 41
PIH_Naked 1.00 0.82 26 26 64 98 41 1.67 0.93 35 21 72 97 41
POG 1.58 1.05 41 26 66 59 41 1.92 1.15 48 25 70 62 41
POI 2.46 1.50 69 28 76 61 41 2.65 1.51 69 26 77 66 41
P/S_I_Love_You 3.00 1.65 66 22 77 63 41 2.50 1.45 55 22 76 65 41
Quality_Earnings 2.40 1.41 60 25 72 58 41 2.25 1.39 45 20 71 63 41
S&P_Peg 1.95 1.24 37 19 68 33 41 1.80 1.09 27 15 66 43 41
Shrinkage 1.29 0.74 22 17 71 22 41 1.73 0.87 26 15 75 28 41
Silver_Parachute 1.46 0.99 35 24 70 44 41 1.24 0.86 21 17 64 53 41
Small_Value 1.44 0.97 36 25 66 44 41 1.70 1.01 39 23 72 46 41
Steady_Growth 1.14 0.74 16 14 61 32 41 1.38 0.59 11 8 61 35 41
Turnarounds 1.82 1.15 51 28 73 90 41 1.50 0.99 39 26 69 87 41
Up_5% 2.08 1.18 54 26 74 91 41 2.38 1.52 50 21 74 95 41
Value_at_the_Top 2.05 1.20 43 21 71 66 41 2.32 1.16 44 19 74 73 41
Zweig-26 2.25 1.53 63 28 69 48 41 1.95 1.23 41 21 71 57 41
Current_Ratio_s -0.59 0.33 -23 39 42 55 41 -0.44 0.44 -14 32 40 61 41
Dilution_s -0.26 0.42 -9 35 47 41 41 -0.67 0.32 -14 21 40 50 41
Negative_FCF_s -0.71 0.29 -27 38 41 86 41 -0.93 0.24 -26 28 36 89 41
The_Mirror_s -0.91 0.26 -29 32 36 82 41 -1.08 0.18 -28 26 35 84 41

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Syntax highlight code surrounded by the {syntaxhighlighter SPEC}...{/syntaxhighlighter} tags, where SPEC is a Syntaxhighlighter options string or class="OPTIONS" [title="the title"].
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd> <h2> <h3>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
CAPTCHA
This question is for testing whether you are a human visitor and to prevent automated spam submissions.