Post New

As a nervous owner of NVDA I have been wondering about selling before the earnings report. My record on previous occasions has not been good. Holding sometimes when selling would have been better and selling when holding would have been better.

Using GTR1, I have tried to see on average the best course of action for NAS100 stocks. The time between earnings reports is not constant, so I choose to sell if it has been greater than 82 days since the last report and by after 3 days.

Buy and hold all 100 NAS100 stocks equal weight since 20000403 (the date GTR! has the first data on earnings date).

CAGR 9.3%
SAWR 4.15%

Selling if the earnings date was more than 82 days ago and buying 3 days afterwards.

CAGR 7.36%
SAWR 3.41%

I guess I will hold NVDA through the earnings day.

https://gtr1.net/2013/?h1r0::linear%281,ord%281%29...


Aussi

The NAS100 has been a winner, and less rebalancing had higher CAGR (with no friction, 40 years from 19850201 to 20241113):

      Screen        CAGR  GSD
NazNoRebalHold1day  15.8  26    hold 1 day
    NazNoRebal      14.5  27   hold 21 days
  NazDecadeRebal    14.7  26   hold 21 days
  Naz5yearRebal     13.1  26   hold 21 days
  NazAnnualRebal    12.2  26   hold 21 days
  NazMonthlyRebal   11.6  27   hold 21 days

The best rule was: never sell. Maybe sell for some specific reason, but in general, selling lost CAGR. The best CAGR was buying a NAS100 stock on the first day it appeared on the list. Waiting a month had slightly lower CAGR.

Selling based on the announcement date (with no friction, 25 years from 20000403 to 20241113):

              Screen                CAGR  GSD  MDD  Sharpe  AT   Depth
            NazNoRebal              8.1   27   -75   0.38   0.2   100
          NazMonthlyRebal           6.9   28   -80   0.33   0.6   100
   NextAnnouncementDateIsCloser     8.7   29   -76   0.39    8     53
   LastAnnouncementDateIsCloser     5.2   27   -80   0.26    7     53
NextAnnouncementDateIsCloser_top30  9.3   30   -74   0.40   12     33
LastAnnouncementDateIsCloser_btm4   13.4  30   -71   0.55   34     6

These screens use [qsdays] which is days since the latest quarterly announcement.

{NextAnnouncementDateIsCloser} picks the highest 50 [qsdays]. This includes stocks that will announce earnings in a few days. CAGR is about the same as {NazNoRebal}.

{LastAnnouncementDateIsCloser} picks the lowest 50 [qsdays]. This includes stocks that have recently announced earnings. CAGR is lower than {NazNoRebal} and {NazMonthlyRebal}.

{NextAnnouncementDateIsCloser_top30} picks the highest 30 [qsdays]. This is a tuned screen, with the magic number 30 resulting in the highest CAGR. The screen holds the 30 stocks that announced earnings longest ago. CAGR is slightly higher than {NazNoRebal}.

{LastAnnouncementDateIsCloser_btm4} picks the lowest 4 [qsdays]. This is a tuned screen, with the magic number 4 resulting in the highest Sharpe ratio. The screen holds the 4 stocks that have most recently announced earnings. CAGR is higher than {NazNoRebal}.

https://gtr1.net/2013/?~NazNoRebal:h21r0::nas100.a...
https://gtr1.net/2013/?~NazMonthlyRebal:h21::nas10...
https://gtr1.net/2013/?~NextAnnouncementDateIsClos...
https://gtr1.net/2013/?~LastAnnouncementDateIsClos...
https://gtr1.net/2013/?~NextAnnouncementDateIsClos...
https://gtr1.net/2013/?~LastAnnouncementDateIsClos...


===== Some notes on the date fields =====
There are period-end dates and announcement dates.

>>>The field I used:
Create [qsdays]: Days since latest quarterly announcement
qsdays: minus(ord(1),qsdate.s)


>>> Period-end dates are:

In SIP, DATE_EQ0 is the end date of the current quarter (not yet filed). For example, November 16th DATE_EQ0 is:
20241231 for AAPL
20241031 for NVDA

SIP field names for quarter ends:
PEREND_Q1 Ending date Q1
DATE_EQ0 Date--Est EQ0 fiscal quarter end

gtr1 field names for quarter ends:
perendq1.s [SI Q1 Ending Date as Ordinal]
dateeq0.s [SI Latest Quarterly EPS Date as Ordinal]


>>> Announcement dates are [field, description, default, start date]:

qsdate.s [SI Quarterly Surprise Date as Ordinal] -9999 20000331
The date on which the latest quarterly announcement was made by a company.

repdtq0.s [SI EPS Est Q0-Exp Rep Date] -999999 20120629
The day on which a company is expected to announce its quarterly earnings for the current fiscal quarter.

Picking the 4 stocks with the most recent earnings announcement works for Nas100, S&P500, and S&P1500. All had higher CAGR, but also extremely high turnover, with about 100% turnover every market day with the S&P 500. Results from 20000403 to 20241113 (25 years):

      Screen        CAGR  GSD  MDD  Sharpe  AT  Depth
  RecentAnnounNaz    13   30   -71   0.55   34    6
RecentAnnounSP500    16   25   -42   0.70   43    11
RecentAnnounSP1500   16   31   -59   0.64   46    25
 SP500EqualWeight    10   21   -60   0.52   0    499
SP1500EqualWeight    11   24   -60   0.51   0   1499

Results from 20141111 to 20241113 (10 years):

      Screen        CAGR  GSD  MDD  Sharpe  AT  Depth
  RecentAnnounNaz    18   24   -38   0.79   36    6
RecentAnnounSP500    17   25   -38   0.71   46    9
RecentAnnounSP1500   13   31   -53   0.53   50    16
 SP500EqualWeight    11   19   -39   0.60   0    500
SP1500EqualWeight    10   23   -42   0.51   0   1499

The excess annual returns are trending down, with {RecentAnnounSP500} having poor returns the last 3 years.

year  recentAnn  SP500ew  diff
2024      5        18     -13
2023      17       14       3
2022     -13       -11     -1
2021      41       30      11
2020      32       12      20
2019      33       30       4
2018      6        -8      14
2017      14       18      -4
2016      34       15      19
2015      1        -2       4
2014      19       14       4
2013      38       36       2
2012      26       17       8
2011      19        0      19
2010      49       22      27
2009      34       46     -12
2008     -29       -40     10
2007      -2        1      -4
2006      19       16       3
2005      -3        8     -11
2004      13       17      -5
2003      36       42      -6
2002     -23       -18     -5
2001      31        0      31
2000      48       10      38
 
avg       18       12       6
 sd       21       19      14
cagr      16       10
 >0      80%       72%    64%

https://gtr1.net/2013/?~SP500EqualWeight:h63f0.4::...(class.a,permco.a,step1)et1
https://gtr1.net/2013/?~SP1500EqualWeight:h63f0.4:...
https://gtr1.net/2013/?~RecentAnnounSP500:h1r0::sp...
https://gtr1.net/2013/?~RecentAnnounSP1500:h1r0::s...
https://gtr1.net/2013/?~RecentAnnounNaz:h1r0::nas1...

Tom Bowley over at Earnings Beats / Stockcharts runs a specific quarterly "screen". It's 1 stock each with highest RS & earnings growth from each of 10 top "industries". It's a quarterly hold (mostly) - screen is refreshed 6 weeks after quarterly earnings season starts.

It holds stocks regardless of behavior (unless a real existential event) through earnings season to the next screen turnover.

Which is a bit weird because most of his other trade recs from other screens I've sen in the last 4 years are short term "DON'T" hold through earnings.

It's been volatile and - as we would expect - directionally driven by the market direction. 22 was a tough year. 20 and 21 were great; easy pickings.

And anecdotally but consistently what I've seen is, if a company doesn't BEAT by MORE than expected, the market selloffs are generally ugly. Sometimes, temporary.

The runups in advance of earnings are the opportunity, usually - then selling the week or week before of earnings to protect the gain.

NVDA has been an exception because there is SO MUCH money riding on it.

FC

Holding through the announcement had slightly higher CAGR than no rebalancing, but also higher GSD (From 20120702 To 20241113, no friction):

      Screen        CAGR  GSD  MDD  Sharpe  AT  Depth
 ThruAnnounNas100   20.3  29   -42   0.79   21    6
  Nas100NoRebal     19.6  21   -37   0.99   0    100
Nas100MonthlyRebal  16.1  19   -31   0.89   1    100

{ThruAnnounNas100} buys the stocks with the soonest expected report date, and holds for 10 days. For example, at times buying the day before earnings and holding for 10 days.

Define {ThruAnnounNas100}
step0: [NASDAQ 100 Member; lag=1 days] == 1
step1: [SI EPS Est Q0-Exp Rep Date; lag=1 days] > 0
step2: [SI EPS Est Q0-Exp Rep Date; lag=1 days] Bottom 4; Cash When Fewer
Holding period = 10 mkt days
Rebalance to equal weight among new positions only
All trades at market close.
https://gtr1.net/2013/?~ThruAnnounNas100:h10r0::na...

Note that repdtq0.s is not an Ordinal (e.g. ord:1 is 26109, and NVDA repdtq0.s is 20241120).

quarter ends:
perendq1.s [SI Q1 Ending Date as Ordinal]
dateeq0.s [SI Latest Quarterly EPS Date as Ordinal]

announcement dates:
qsdate.s [SI Quarterly Surprise Date as Ordinal]
repdtq0.s [SI EPS Est Q0-Exp Rep Date]

It seems to me that all these variations of holding period on the NAS100 are insignificant data mining. The elephant in the room would be a cap weighted NAS100 holding, like QQQ. I have a hunch that QQQ would have beaten an equal weighted NAS100 by knockout in a backtest, with 1/100th of the effort, which is of course not a guarantee of future returns.

Elan

The elephant in the room would be a cap weighted NAS100 holding, like QQQ. I have a hunch that QQQ would have beaten an equal weighted NAS100 by knockout in a backtest, with 1/100th of the effort, which is of course not a guarantee of future returns.


Two minor comments (from an admitted QQQE fan)

(1)
Without a doubt, QQQ has beat QQQE for a while, the last 8 or so years--and by a lot. I just don't think that observation has any predictive power, as it's really a wager on a statistically insignificant number of specific firms and their idiosyncratic futures. I think of the returns of QQQ as the returns of QQQE which has some statistical breadth to it, plus or minus a big random number.

(2)
Geek note, QQQ isn't cap weighted. It has its own unique weighting formula, and the difference is material.

Jim

Without a doubt, QQQ has beat QQQE for a while, the last 8 or so years--and by a lot. I just don't think that observation has any predictive power, as it's really a wager on a statistically insignificant number of specific firms and their idiosyncratic futures. I think of the returns of QQQ as the returns of QQQE which has some statistical breadth to it, plus or minus a big random number.

I think the choice between QQQ and QQQE is a market timing bet. IN strong bull markets QQQ will be far ahead, particularly when the bullishness is concentrated in tech companies which it has been for at least 30 years and counting. When there's a bear market or a bubble popping, QQQE will be ahead. I don't think it's random at all.

Elan

When there's a bear market or a bubble popping, QQQE will be ahead. I don't think it's random at all.

QQQE will be ahead in the sense that it loses less than QQQ. Both will lose.
Which begs the question, if you know that you should switch from QQQ to QQQE, why wouldn't you just go to cash?

Geek note, QQQ isn't cap weighted. It has its own unique weighting formula, and the difference is material.

I think QQQ weighting matches NDX weighting within fairly tight limits which are set out in the prospectus.

The weighting in the index is modified to reduce the weighting of a stock if it exceeds a certain limit, or a group of stocks weighting can be reduced if their aggregate weighting exceeds a certain amount.

The adjustments push (slightly)QQQ towards QQQE weighting



https://connect.rightprospectus.com/Invesco/TVT/46...

Rebalancing of the Index
The Index is calculated using a “modified market capitalization-weighted”
methodology, which is a hybrid between equal weighting and conventional
capitalization weighting. This methodology is expected to: (1) retain in general the
economic attributes of capitalization weighting; (2) promote portfolio weight
diversification (thereby limiting domination of the Index by a few large stocks);
(3) reduce Index performance distortion by preserving the capitalization ranking of
companies; and (4) reduce market impact on the smallest Index Securities from
necessary weight rebalancings.

Aussi

Using GTR1 and ^N1T and ^N1TE which are similar to but not the same as QQQ and QQQE due to NDX modifying the weights as discussed up thread, I looked at the performance of the two from 1972.

As Rayvt mentioned, in a down market, ^N1TE usually looses less, but should be in cash.

^N1TE did outperform ^N1T from the bottom in 2002 to the top in 2008 (quite substantially, total return 3.4 vs 2.4). So for 6 years, ^N1TE performed better, but generally for the other 46 years it was better to be in the cap weighted (^N1T) or cash.

I have posted a graph here. https://discussion.fool.com/t/comparison-of-qqq-an...

Aussi

Aussi wrote: Using GTR1 and ^N1T and ^N1TE which are similar to but not the same as QQQ and QQQE due to NDX modifying the weights as discussed up thread,

I took a few moments to compare the returns of the two using one of my simpler trading systems. Here is a bit of a definition of terms:

C = Cash
B = Bull
EB = Extremely Bullish

What is interesting to note is that as stated above market cap weighted has been winning consistently from 2016, but that has not always been the case. What is surprising is that Equal Weighted wins quite often in the middle of a bear market like 2000-2002 and after the end of a major bear market, i.e., 2009. The drop in 2020 had almost exactly the same loss for both.

It is important to note that the winning gains of Market Cap recently are very similar to the winning gains of Market Cap from 1992 to 2000.

Date        Sig  CAGR to Date  ^N1T     ^N1TE    Winner
1/1/1990    B    17.75%        9.16%    9.22%    ^N1TE
7/16/1990   C    14.74%                          
10/10/1990  EB   47.47%        174.03%  155.90%  ^N1T
10/12/1992  B    27.55%        260.35%  112.66%  ^N1T
10/7/1998   EB   47.06%        537.30%  483.05%  ^N1T
3/21/2000   C    42.68%                          
4/4/2001    EB   49.42%        73.32%   80.92%   ^N1TE
5/21/2001   C    47.86%                          
9/21/2001   EB   54.27%        71.74%   79.15%   ^N1TE
11/14/2001  C    49.89%                          
10/7/2002   EB   48.79%        182.52%  345.86%  ^N1TE
3/31/2006   C    40.82%                          
3/6/2009    EB   44.68%        220.09%  283.64%  ^N1TE
3/7/2011    C    43.28%                          
10/4/2011   EB   43.52%        112.38%  116.42%  ^N1TE
10/4/2013   B    39.97%        27.51%   15.07%   ^N1T
2/9/2016    EB   40.45%        39.46%   48.70%   ^N1TE
12/1/2016   B    38.17%        28.46%   22.04%   ^N1T
12/24/2018  EB   39.91%        129.18%  91.46%   ^N1T
2/10/2020   B    38.54%        -22.92%  -22.95%  ^N1T
3/23/2020   EB   40.34%        210.57%  171.04%  ^N1T
3/23/2022   C    39.60%                          
10/12/2022  EB   41.00%        198.77%  109.73%  ^N1T
7/16/2024   C    40.59%                          
11/18/2024                                       

https://zeelotes.org/NDX_Mkt-Cap_or_Eq-Wgt.PNG

Using GTR1 and ^N1T and ^N1TE...

A perhaps stupid question, but is there a list anywhere of indexes that GTR1 recognizes? There are quite a few that it does not, though it's also possible that I just can't figure out how to enter them. My main interest would be in the S&P 400 and 600.

Baltassar

Baltassar asked: A perhaps stupid question, but is there a list anywhere of indexes that GTR1 recognizes? There are quite a few that it does not, though it's also possible that I just can't figure out how to enter them. My main interest would be in the S&P 400 and 600.

Start Date  Symbol  Index
12/31/1925  ^S4T    S&P 400 - Tot Ret - Mkt Cap Wt
12/31/1925  ^S4TE   S&P 400 - Tot Ret - Equal Wt
12/31/1925  ^S4TL2  S&P 400 - Tot Ret - Mkt Cap Wt - Long 2x
12/31/1925  ^S4TL3  S&P 400 - Tot Ret - Mkt Cap Wt - Long 3x
12/31/1925  ^S4TS1  S&P 400 - Tot Ret - Mkt Cap Wt - Short 1x
12/31/1925  ^S4TS2  S&P 400 - Tot Ret - Mkt Cap Wt - Short 2x
12/31/1925  ^S4TS3  S&P 400 - Tot Ret - Mkt Cap Wt - Short 3x
12/31/1925  ^S4T    S&P 400 - Tot Ret - Mkt Cap Wt
12/31/1925  ^S4TE   S&P 400 - Tot Ret - Equal Wt
12/31/1925  ^S4TL2  S&P 400 - Tot Ret - Mkt Cap Wt - Long 2x
12/31/1925  ^S4TL3  S&P 400 - Tot Ret - Mkt Cap Wt - Long 3x
12/31/1925  ^S4TS1  S&P 400 - Tot Ret - Mkt Cap Wt - Short 1x
12/31/1925  ^S4TS2  S&P 400 - Tot Ret - Mkt Cap Wt - Short 2x
12/31/1925  ^S4TS3  S&P 400 - Tot Ret - Mkt Cap Wt - Short 3x
12/31/1925  ^S4TS3  S&P 400 - Tot Ret - Mkt Cap Wt - Short 3x
12/31/1925  ^S4TL3  S&P 400 - Tot Ret - Mkt Cap Wt - Long 3x
12/31/1925  ^S4TS1  S&P 400 - Tot Ret - Mkt Cap Wt - Short 1x
12/31/1925  ^S4TS2  S&P 400 - Tot Ret - Mkt Cap Wt - Short 2x
12/31/1925  ^S4T    S&P 400 - Tot Ret - Mkt Cap Wt
12/31/1925  ^S4TE   S&P 400 - Tot Ret - Equal Wt
12/31/1925  ^S4TL2  S&P 400 - Tot Ret - Mkt Cap Wt - Long 2x
12/31/1925  ^S6T    S&P 600 - Tot Ret - Mkt Cap Wt
12/31/1925  ^S6TE   S&P 600 - Tot Ret - Equal Wt
12/31/1925  ^S6TL2  S&P 600 - Tot Ret - Mkt Cap Wt - Long 2x
12/31/1925  ^S6TL3  S&P 600 - Tot Ret - Mkt Cap Wt - Long 3x
12/31/1925  ^S6TS1  S&P 600 - Tot Ret - Mkt Cap Wt - Short 1x
12/31/1925  ^S6TS2  S&P 600 - Tot Ret - Mkt Cap Wt - Short 2x
12/31/1925  ^S6TS3  S&P 600 - Tot Ret - Mkt Cap Wt - Short 3x
12/31/1925  ^S6T    S&P 600 - Tot Ret - Mkt Cap Wt
12/31/1925  ^S6TE   S&P 600 - Tot Ret - Equal Wt
12/31/1925  ^S6TL2  S&P 600 - Tot Ret - Mkt Cap Wt - Long 2x
12/31/1925  ^S6TL3  S&P 600 - Tot Ret - Mkt Cap Wt - Long 3x
12/31/1925  ^S6TS1  S&P 600 - Tot Ret - Mkt Cap Wt - Short 1x
12/31/1925  ^S6TS2  S&P 600 - Tot Ret - Mkt Cap Wt - Short 2x
12/31/1925  ^S6TS3  S&P 600 - Tot Ret - Mkt Cap Wt - Short 3x
12/31/1925  ^S6TS3  S&P 600 - Tot Ret - Mkt Cap Wt - Short 3x
12/31/1925  ^S6TS1  S&P 600 - Tot Ret - Mkt Cap Wt - Short 1x
12/31/1925  ^S6TS2  S&P 600 - Tot Ret - Mkt Cap Wt - Short 2x
12/31/1925  ^S6TL3  S&P 600 - Tot Ret - Mkt Cap Wt - Long 3x
12/31/1925  ^S6TL2  S&P 600 - Tot Ret - Mkt Cap Wt - Long 2x
12/31/1925  ^S6T    S&P 600 - Tot Ret - Mkt Cap Wt
12/31/1925  ^S6TE   S&P 600 - Tot Ret - Equal Wt

Wow! Thanks so much!!

But now I have to ask, where should I have looked?

Baltassar

I think the choice between QQQ and QQQE is a market timing bet. IN strong bull markets QQQ will be far ahead, particularly when the bullishness is concentrated in tech companies which it has been for at least 30 years and counting. When there's a bear market or a bubble popping, QQQE will be ahead.

Why would that necessarily be the case? For example, the larger companies have deeper pockets to weather downturns and lots more cash.

Here are the top and bottom holdings in QQQ along with their sectors; both are heavy in 'information technology'. Is biotech more or less volatile?

AAPL  Info Tech    MRNA  Health Care
NVDA  Info Tech    ARM   Info Tech
MSFT  Info Tech    SMCI  Info Tech
AMZN  Cons Disc    ILMN  Health Care
AVGO  Info Tech    BIIB  Health Care
META  Comm Serv    GFS   Info Tech
TSLA  Cons Disc    CDW   Info Tech
COST  Cons Stap    MDB   Info Tech
NFLX  Comm Serv    WBD   Comm Serv
GOOG  Comm Serv    ON    Info Tech

DB2

Post New

US Policy FAQ

---

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