Post New

There is much to be said for simplicity.
Simple screen, simple post

Value Line 1700 universe, no crap filters.

The screen:
Sales Growth 5-Year top 300
Stock Price / Low Price 52-Week top 10

That's it.

Really zooms in a strong market, especially during rebounds from a bear.
But also quite decent on average at other times, giving a huge overall return in backtest with risk metrics for a 10-stock portfolio only a little worse than the S&P.

Monthly, 0.4% friction

      Year    Screen      S&P      Diff
      1997      31.3     27.9      +3.4
      1998      86.1     34.5     +51.6
      1999     172.5     18.4    +154.1
      2000      19.3    -10.8     +30.1
      2001       7.3     -8.0     +15.3
      2002     -45.9    -18.9     -27.0
      2003     154.5     23.0    +131.5
      2004      32.0      8.9     +23.1
      2005      14.2      7.5      +6.7
      2006      28.1     13.8     +14.3
      2007      -2.4      1.9      -4.3
      2008     -29.9    -32.9      +3.0
      2009     144.0     25.5    +118.5
      2010      25.5     14.2     +11.4
      2011      -1.0      2.5      -3.5
      2012      -4.0     17.1     -21.1
      2013     106.8     27.6     +79.2
      2014      23.8     12.9     +10.9
      2015       9.9      1.9      +7.9
      2016      28.1     14.4     +13.7
      2017      36.8     21.8     +15.0
      2018      -8.1     -3.5      -4.6
      2019      12.4     29.9     -17.5
      2020     127.3     16.9    +110.4
      2021      36.5     30.5      +6.0
      2022     -10.9    -19.0      +8.1
      2023      18.4     17.5      +0.9  (to June, not annualized)
 
26.5 years      28.8      9.0     +19.9

DDD3 risk      13.3%     11.0%

This backtest is done using the old backtest.org trade dates.

I wouldn't put all my money in this, but it might add a little leavening to the portfolio loaf.
It certainly isn't hard to figure out the picks.
But if doing the momentum step yourself from scratch, recall that VL's "Low Price 52-Week" field is calculated only on data to the end of the prior calendar month, and is available only a couple/few days later.

Jim

Hey Jim. Love your posts as always.

Would you mind confirming the results of the screen above with the slightly different screen posted (post number 2926) from the Brk Board:
Thanks!



"It can't get much simpler than this:
* Start with the ~1700 stocks in the Value Line database.
* Find the 40 stocks with the highest reported five-year rate of sales growth per share. (this field is updated only annually for each stock)
* Of those 40, buy equal dollar amounts of the 10 stocks which are currently furthest above their respective 52 week low prices. (measured one day before your trading)
* Hold those 10 stocks for a month, repeat.

Unsurprisingly, this likes exuberant markets, including bungee rebounds.
The surprising thing is that it does not that badly on average the rest of the time, too.
Final result 28.8%/year 1997 to date versus 9.0% for the S&P 500, at least in backtest.
On relatively sane risk metrics (size and frequency of rolling years having a shortfall below a reasonable return), this 10 stock portfolio is only a tad riskier than the S&P 500.
Mixing a little of this yeast into the "loaf" of your portfolio wouldn't be so crazy.
Come to the dark side!

Figures after trading costs, includes dividends, no provision for tax.
Year Screen S&P
1997 43.3 27.9
1998 86.1 34.5
1999 172.5 18.4
2000 19.3 -10.8
2001 7.3 -8.0
2002 -45.9 -18.9
2003 154.5 23.0
2004 32.0 8.9
2005 14.2 7.5
2006 28.1 13.8
2007 -2.4 1.9
2008 -29.9 -32.9
2009 144.0 25.5
2010 25.5 14.2
2011 -1.0 2.5
2012 -4.0 17.1
2013 106.8 27.6
2014 23.8 12.9
2015 9.9 1.9
2016 28.1 14.4
2017 36.8 21.8
2018 -8.1 -3.5
2019 12.4 29.9
2020 127.3 16.9
2021 36.5 30.5
2022 -10.9 -19.0
2023 18.4 17.5 (half year not annualized)"

Same screen.
I just thought I would repost it to the more suitable board.

The only difference is the typo in this thread where I describe the steps.
It should say "top 40" not "top 300". Not sure how that got in there! Sorry about that~~

Jim

No worries man! Love the screen. Though probably will wait awhile before dipping my toe in the water.

I have another question as well. What was the worst intra-year drawdown? Just wondering what that 2002 beatdown looked like at its absolute worst point (end of month anyway...not that it couldn't do even worse than that still;)

-John

What was the worst intra-year drawdown?

I don't have daily portfolio balances for it.
But looking only at balances on monthly snapshots, worst down-from-all-time-high drawdown was -62.7% and maximum time to get a new all time high was 34 months ending August 2003.

For drawdown-from-highest-point-in-prior-12-months, it was similar at -59% hit in October 2002.


The equivalent figures for the S&P 500 are -53.1% (from all time high) and -49.1% (from trailing 12 month high)
Not all that much better.
I didn't check the Nasdaq on those dates!

Jim

But looking only at balances on monthly snapshots, worst down-from-all-time-high drawdown was -62.7% and maximum time to get a new all time high was 34 months ending August 2003.
For drawdown-from-highest-point-in-prior-12-months, it was similar at -59% hit in October 2002.

The equivalent figures for the S&P 500 are -53.1% (from all time high) and -49.1% (from trailing 12 month high)
Not all that much better.
I didn't check the Nasdaq on those dates!

Curiosity got the better of me.
The equivalent figures are -82.0% and -65.0%.

So, on that count the screen is quite a bit better than QQQ but still somewhat worse than SPY.

Jim

From 1997 to 2022, the OP screen had a 28% CAGR and -63% MDD. A gtr1 version {SG_PriceLow} had 19% CAGR and -83% MDD. (The CAGR range was 14 to 25 for the 21 different start dates). (The gtr1 version does not use VL data, and that might account for the different results.)

The OP screen looks for very high Sales Growth, and then does a final sort using (Price / Low Price). The final sort looks for stocks far above their recent low price. Recent high price is not used.

An alternative screen {SG_HighToLow} looks for stocks highest in the range 52-week High Price to 52-week Low Price. This finds stocks near the High Price and far from the Low Price. High [HtoL] where:

LP = [Lowest Price over 252 days]
HP = [Highest Price over 252 days]
[HtoL] = (P - LP) / (HP - LP)

[HtoL] is 1 when P=HP
[HtoL] is 0 when P=LP

{SG_PriceLow} and {SG_HighToLow} had similar CAGR, but {SG_HighToLow} has lower GSD and MDD.

        Screen         CAGR  SAWR  GSD  MDD  Sharpe  AT    From       To     SDcg
SalesGrowth_PriceLow   18.8  8.1   41   -83   0.62   4   19961231  20221230  2.7
SalesGrowth_HighToLow  18.6  10.8  30   -60   0.72   5   19961231  20221230  1.1
 
  SP500MktCapWeight    8.6   5.0   19   -55   0.46   0   19961231  20221230  0.0
  SP500EqualWeight     10.2  7.7   21   -60   0.52   0   19961231  20221230  0.1
  SP1500EqualWeight    10.6  8.2   24   -60   0.50   0   19961231  20221230  0.2

          return       return         return       return   return     win          win             win         win
 Year   OP_Screen  {SG_PriceLow}  {SG_HighToLow}  SP1500ew    S&P   OP_Screen  {SG_PriceLow}  {SG_HighToLow}  SP1500ew
 1997       31           -6             -1           27       28        1            0              0             0
 1998       86           49             67            6       35        1            1              1             0
 1999      173          171             138          11       18        1            1              1             0
 2000       19           12             -4           14       -11       1            1              1             1
 2001       7           -30             -4           10       -8        1            0              1             1
 2002      -46          -53             -22         -15       -19       0            0              0             1
 2003      155           78             43           44       23        1            1              1             1
 2004       32           25             28           19       9         1            1              1             1
 2005       14           16             23            8       8         1            1              1             1
 2006       28           19             19           16       14        1            1              1             1
 2007       -2           45             48           -1       2         0            1              1             0
 2008      -30          -30             -13         -36       -33       1            1              1             0
 2009      144           68             40           46       26        1            1              1             1
 2010       26           46             26           26       14        1            1              1             1
 2011       -1           30             9            -1       3         0            1              1             0
 2012       -4           32             45           18       17        0            1              1             1
 2013      107           79             56           39       28        1            1              1             1
 2014       24           19             27            9       13        1            1              1             0
 2015       10           -5             -7           -5       2         1            0              0             0
 2016       28          -16             -7           24       14        1            0              0             1
 2017       37           63             42           14       22        1            1              1             0
 2018       -8          -22             -18         -10       -4        0            0              0             0
 2019       12           50             40           25       30        0            1              1             0
 2020      127           92             45           14       17        1            1              1             0
 2021       37           14             -4           31       31        1            0              0             1
 2022      -11          -28             -17         -13       -19       1            0              1             1
 
  avg       38           28             23           12       10       77%          69%             77%         54%
median      24           25             26           14       14
  max      173          171             138          46       31
  min      -46          -53             -22         -36       -33
  sd        58           48             35           19       17
 cagr      28%          19%             19%         11%       8%

http://gtr1.net/2013/?~SalesGrowth_PriceLow:h21f0....
http://gtr1.net/2013/?~SalesGrowth_HighToLow:h21f0...

The March 10th picks for {SG_HighToLow} were:

Ticker  styp.a  [MktCap]  [SalesY6]   [SG5]   [SharesY6]  [ShrG5]  [SShG5]  gprc:1  lgprc:1,252  hgprc:1,252  [HtoL52c]  [PL52c]  rankPL
 SRPT     11     12,662     5.40     129.98     48.70       1.67    77.89    4.33       1.89         4.70       0.87       2.30     4
 MELI     11     58,703     844.40     8.37     44.16       1.13     7.42   47.59      24.96        51.54       0.85       1.91     9
  PR      11      3,111     98.80     10.42     165.68      1.70     6.15    1.09       0.53         1.20       0.83       2.03     7
  YPF     31      1,840    1075.70     6.05     391.50      1.00     6.03    2.29       0.62         2.73       0.79       3.66     2
 RVNC     11      2,371     0.30     259.33     28.11       2.40   108.00    1.07       0.43         1.32       0.72       2.49     3
 FOLD     11      3,234     5.00      61.10     134.40      2.02    30.26    0.89       0.48         1.05       0.71       1.84     10
 EXAS     11     11,048     266.00     7.84     115.68      1.52     5.14    4.94       2.42         6.08       0.69       2.04     6
  WTM     12      3,552     157.70     5.99      5.01       0.61     9.76   97.98      73.70       109.82       0.67       1.33     26
 TGTX     11      2,061     0.20      33.50     49.04       2.70    12.43    0.02       0.00         0.02       0.66       3.76     1
  VMW     11     50,047     512.00    25.10     408.63      1.03    24.45    3.74       2.97         4.19       0.63       1.26     30
 
 BGNE     31     19,546     1.10     1069.36    31.05       2.99   357.83    7.65       4.28         9.71       0.62       1.79     12
 PRTA     12      2,522     1.10     182.36     34.35       1.29   141.64   10.29       4.88        13.94       0.60       2.11     5
  LNG     11     38,351    5601.00     5.97     233.10      1.08     5.54   14.83      11.69        17.21       0.57       1.27     28
 SNDX     11      1,522     1.20     116.42     14.62       3.56    32.69    1.86       1.14         2.42       0.56       1.63     14
 CELH     11      6,367     22.80     13.79     38.57       1.91     7.21    2.19       1.09         3.12       0.54       2.01     8
 ASND     31      5,322     1.60      33.88     33.63       1.67    20.31    5.65       4.15         7.02       0.52       1.36     25

{SG_PriceLow} picks were mostly the same, with only 2 out of 10 different:

Ticker  gprc:1  lgprc:1,252  hgprc:1,252  [HtoL52c]  [PL52c]  rankPL  rankHL  {SG_PriceLow}  {SG_HighToLow}  HP/LP
  WTM   97.98      73.70       109.82       0.67       1.33     26      8           no             yes       1.49
  VMW    3.74       2.97         4.19       0.63       1.26     30      10          no             yes       1.41
 
 PRTA   10.29       4.88        13.94       0.60       2.11     5       13         yes             no        2.86
 CELH    2.19       1.09         3.12       0.54       2.01     8       16         yes             no        2.86

{SG_PriceLow} picks more volatile stocks (stocks with a larger low to high range).

From 1997 to 2022, the OP screen had a 28% CAGR and -63% MDD. A gtr1 version {SG_PriceLow} had 19% CAGR and -83% MDD. ..

Thanks for creating an alternative version. Nice work

Two comments:
* As mentioned, my drawdown figures were using balances only at monthly trading dates, so intra-month drawdowns will necessarily be worse.

* A different source database of course means a different screen.
It might be worse, or it might be better, but it's different.

For example, I believe the "Sales Growth 5-Year" field at VL is the 3-year average divided by the 3-year-average five years earlier.
(that's how they do their earnings growth figures, though I don't really know about the sales growth figure, maybe it's simply 5 year CAGR of sales per share)
Or, the VL database isn't just the largest by market cap, they do a lot of hand picking for the smaller stocks.
And so forth

Jim

Post New

Fidelity National Information Services FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.