Studies - Posted on Friday, January 1, 2010, 11:32 PM GMT +1

12 Comments


Jan Friday 1

How To Make Millions (in %) Trading The SPYDER – Seasonalities (I)

As promised this will be the first in a series of posts about my my step-by-step approach, my initial basic parameters, and respective results (findings) with respect to a market model (a concise mathematical formula which would stand the test of time forecasting the next sessions S&P 500‘ / the SPY‘s performance on the close with maximum accuracy) trying to make greed and fear as the never changing driving forces behind market movements somewhat ‘quantifiable’ (and of course in order to set up a hopefully profitable end-of-day trading strategy as well).

The first part – titled Seasonalities – will deal with potential edges provided on the long and/or short side of the market due to periodical calendar events like the last session(s)/the first session(s) of a month, exchange holidays, FOMC meetings (FOMC announcement sessions), Government’s Job Report (regularly on the first Friday of a month), option expiration (on the third Friday of a month), the days of a week, the week of a year or the month of a year (but both will – at least for the time being – not be part of the evaluation process due to the fact that the model’s time horizon is the next session only), in order to evaluate if – and to what extent- these periodical events had – and probably will have in the future – a (significant) impact on the direction and magnitude of the next session’s change.

Please take into account that the dates presented below (e.g. the calendar day of the month, the day of the week, an FOMC announcement session, among others) and their respective performance figures do NOT reflect the index’ performance on the the session itself, but the next session’s performance instead, means if -and to what extent- the respective date (periodical event) would’ve provided a favorable opportunity for going long or short on the close targeting a higher/lower close on the then following session.

xx

1. The (calendar) Day of the Month

For those of you being regular readers of any kind of quant blogs, you might be familiar with the stock market adage that especially the last and the first sessions of a month – from a historically perspective – show a positive bias in comparison to any other day of the month.

Table I below shows the S&P 500‘ historical performance (Cumulative Returns, Profit Factor, Win/Loss Ratio and the t-score vs. the Index) concerning the calendar days of the month (1 – 31), assumed one would’ve bought the S&P 500 on close of the respective calendar day and sold the index on close of the then following session,  for the time period between 01/01/1940 and 12/31/2009 (70 years).

Not surprisingly (and confirming the market adage), and from a long-term historical perspective, the last four and first two calendar days of a month provided the most favorable opportunities for going long on the close (day 29, 30, 31, 1 and 2 with an associated t-score exceeding the +1.645 mark for statistical significance, means there is a low probability that this ‘abnormality’ occurred by chance only).

But only calendar days 30 and 31 stood the test of time.  Table II below shows the respective S&P 500‘ historical performance, now for the time period between 01/01/1990 and 12/31/2009 (the last 20 years).

Only calendar day 30 (and 31 is close) has been able not only to keep, but to improve it’s respective performance figures (Profit Factor, Win/Loss Ratio, t-score).

First conclusion with respect to the market model: The model will take a long position on calendar day 30 or day 31, depending on which one represents the last business day/session of a month (for the time being not depending on any additional condition).

xx

2. The Day of the Week

On step 2 we’ll check if – and to what extent – the day of the week might provide a favorable opportunity on any side of the market.

Table III below shows the S&P 500‘ historical performance (Cumulative Returns, Profit Factor, Win/Loss Ratio and the t-score vs. the Index) concerning the day of the week (2 = Monday, 3 = Tuesday, …, 6 = Friday), assumed one would’ve bought the S&P 500 on close of the respective weekday and sold the index on close of the then following session, for the time period between 01/01/1940 and 12/31/2009 (70 years).

Interestingly – from a long-term historical perspective – Tuesdays have been the most favorable weekday to go long on close targeting a higher close on the then following session (regularly Wednesday if not being an exchange holiday), but the respective Profit Factor and Win/Loss ratio are nothing to write home about. It seems that Tuesday’s might have profited from some extraordinary positive market movements on Wednesdays (in contrast to calendar day 30 which shows a far superior profit Factor and Win/Loss ratio).

In addition, looking a the recent history (the last 20 years), any potential edge completely disappeared.

Second conclusion with respect to the market model: The model will NOT take into account the day of the week.

xx

3. FOMC Meetings (scheduled) and FOMC announcement sessions

The Fed announces its policy decision (regularly whether to change the fed funds target rate) at the end of each FOMC meeting (FOMC announcement session). The FOMC announcement session shows a (highly) favorable positive bias (going long on close of the pre-FOMC announcement session), and it pays as well to go long on close of the FOMC announcement session assumed the index didn’t close above the +0.50% mark.

Table IV below shows the S&P 500‘ historical performance (Cumulative Returns, Profit Factor, Win/Loss Ratio and the t-score vs. the Index) concerning the pre-FOMC announcement session (setup 1), the FOMC announcement session assumed the SPX didn’t close up greater than +0.50% (setup 2), the combination of setup 1 and 2 (OR = going long on close of any of both session, setup 3), and the FOMC announcement session assumed the SPX closed up greater than +0.50% (setup 4), for the time period between 01/01/1990 and 12/31/2009 (the last 20 years).

Interesting to note that setup 1 (going long on close of an pre-FOMC announcement session) shows an excellent median winning trade of +0.80%, significantly higher than the at-any-time winning trade of +0.54%, and a median losing trade of -0.40% only, significantly lower than the at-any-time losing trade (lower index close), while the winning percentage is only slightly above the at-any-time winning percentage. The same applies to setup 3, the combination of going long on a pre-FOMC announcement session and going long on close of the FOMC announcement session if the index didn’t close up greater than +0.50%. But in the event the index did close up greater than +0.50% on an FOMC announcement session, the odds do NOT favor the buy side any longer (setup 4).

Third conclusion with respect to the market model: The model will take a long position on close of a pre-FOMC announcement session as well as on close of an FOMC announcement session if the index doesn’t close up greater than +0.50%.

xx

4. Government’s Job Report Friday

The Government’s Job Report is regularly published on the first Friday of every month, if the Friday itself and/or the preceding Thursday doesn’t fall on a public holiday (e.g. July 4). For simplification purposed I assumed that it would’ve been always released on the first Friday of every month (hopefully averaging out any deviations from the rule). In addition, I don’t know when the Government’s Job Report had been published for the first time, and if it had been published always on a Friday during the last 70 years.

Table V below shows the S&P 500‘ historical performance (Cumulative Returns, Profit Factor, Win/Loss Ratio and the t-score vs. the Index) concerning the session immediately preceding the first Friday of every month (assumed going on on close of the respective session) –setup 1-, and the same session with those occurrences only where the index closed higher the session immediately preceding the first Friday of every month (setup 2), for the time period between 01/01/1940 and 12/31/2009 (70 years).

Interesting to note that a higher close on a session immediately preceding the Government’s Job Report Friday provides a favorable edge on the long side of the market (winning percentage, profit factor, median winning and median losing trade) on Job Report Friday as well (no short-term mean reversion), superior to those odds when not taking into account the direction of the close on a session immediately preceding the Government’s Job Report Friday (trend-following seams more advisable than betting on short-term mean reversion).

Fortunately the edge is provided during the recent 20 years as well (profit factor and winning percentage almost unchanged). Table VI below shows the S&P 500‘ historical performance (Cumulative Returns, Profit Factor, Win/Loss Ratio and the t-score vs. the Index) concerning the session immediately preceding the first Friday of every month (assumed going on on close of the respective session) –setup 1-, and the same session with those occurrences only where the index closed higher the session immediately preceding the first Friday of every month (setup 2), for the time period between 01/01/1990 and 12/31/2009 (the last 20 years).

Fourth conclusion with respect to the market model: The model will take a long position on close of a a session immediately preceding the Government’s Job Report Friday in the event the index closed up.

xx

5. Option Expiration

Expiration day for equity and index options is the Saturday immediately following the third Friday of the expiration month.

Table VII below shows the S&P 500‘ historical performance (Cumulative Returns, Profit Factor, Win/Loss Ratio and the t-score vs. the Index) assumed going long on on close of the session immediately preceding the third Friday of every month (setup 1), going long on close of the third Friday (setup 2) and going long on close of the session immediately following the third Friday (setup 3), for the time period between 01/01/1990 and 12/31/2009 (the last 20 years).

The sessions on or around the third Friday of a month obviously do not provide a favorable opportunity on the long side of the market (especially not for a long entry on close of option expiration Friday itself), but without taking into account any additional condition (like a higher/lower close), no significant edge is provided.

Even with respect to the market’s long-term history (with or without the availability of options contracts), the session on or immediately preceding the third Friday of a month did never provide a favorable opportunity for going long. Table VIII below shows the S&P 500‘ historical performance (Cumulative Returns, Profit Factor, Win/Loss Ratio and the t-score vs. the Index) assumed going long on on close of the session immediately preceding the third Friday of every month (setup 1), going long on close of the third Friday (setup 2) and going long on close of the session immediately following the third Friday (setup 3), for the time period between 01/01/1940 and 12/31/2009 (the last 70 years).

______________________

Due to the fact the this posting is getting a bit long in the tooth, it will be continued then dealing with the session around exchange market holidays.

For those interested in, the model’s signals will be posted from January 4, 2010 (next Monday) onward regularly shortly before the close (but even if posted shortly after the close, the SPY is still available for trading with good liquidity and narrow spreads of regularly $0.01) via Twitter Update by following a new Twitter account @Strategy_I . (making further improvements to the model – e.g. optimization of the short side and reducing the chance of overfitting – as well as introducing position sizing, leverage, abnormal market filters, implementing buy/sell stops and probably making the model ‘adaptive‘ will be an ongoing process in 2010)

But please be reminded that past performance is never an indication (leave alone a guarantee) for future performance, and for good reason the CFTA and NFA (COMPLIANCE RULE 2-29: USE OF PROMOTIONAL MATERIAL CONTAINING HYPOTHETICAL PERFORMANCE RESULTS) requires all members (which I’m not) to explicitly reference to limitations, risks, lack of liquidity, among others, and publish a respective disclaimer.

Successful trading,
Frank

________________________________

If you might want to be instantly notified about what’s happening in the markets and at TRADING THE ODDS, I encourage you to subscribe to my RSS Feed or Email Feed, and (or) follow me on Twitter.

xx

Disclaimer: No position in the securities mentioned in this post at time of writing.

The information on this site is provided for statistical and informational purposes only. Nothing herein should be interpreted or regarded as personalized investment advice or to state or imply that past results are an indication of future performance. The author of this website is not a licensed financial advisor and will not accept liability for any loss or damage, including without limitation to, any loss of profit, which may arise directly or indirectly from use of or reliance on the content of this website(s). Under no circumstances does this information represent an advice or recommendation to buy, sell or hold any security.

Add to Technorati Favorites

Comments (12)

 

  1. CarlosR says:

    Hi Frank,

    Great work, as usual, thanks for publishing it.

    I know you have a lot on your plate, but when you have time, you might find some interesting additional ideas for seasonal effects here: http://calendar-effects.behaviouralfinance.net/

    One that caught my eye was this one: “Kamstra, Kramer and Levi (2000) found a daylight saving anomaly. Daylight-saving weekends are typically followed by large negative returns on financial market indices (roughly 200 to 500 percent of the regular weekend effect), and they argue that the effect could be a direct result of changes in sleep patterns. ”

    And there’s lots of others where that one came from. (the entire behaviouralfinance.net web site is full of good leads)

  2. Noam says:

    Great post

    Thank you so much!!

    A suggestion, could you make some seasonality charts on OIL, GOLD, SILVER, GRAINS and the Chemical stocks which related to the Grains (POT, MOS ,DOW)

    Thanks!!!

  3. JP says:

    Hi Frank,

    I ran a test buying on the close of the first Thursday of the month and exiting at the close of the next session. My numbers aren’t quite as good as yours, but they are profitable. The issue that I see is that this signal had a great run from 1997-2000 and hasn’t made a new equity high since. There are two large and prolonged drawdown periods which make this unsuitable for me to trade it.

    So I figure that either I’m running something wrong, or if I’m right that I wanted to let you know. Any feedback is appreciated.

    Cheers,
    JP

    • Jp,

      the ‘Government Job Report’ setup is a) NOT triggered on close of the first Thursday of a month (it is triggered on close of the session preceding the first Friday of a month which makes quite a difference over the course of 20+ years; think about Friday being the 1st day of a month), and b) is only triggered in the event the SPY closed higher the session immediately preceding the first Friday of a month.

      I hope that helps.

      Best,
      Frank

      • JP says:

        Frank,

        You are correct. My testing didn’t account for the Friday on the 1st on a month. I left this out when I first started in order to keep the code simple and then later forgot to come back and fix it. It makes sense the that 1st of the month could significantly impact the results. Appreciate the reply.

        Cheers,
        JP

        • JP says:

          Frank,

          Not to belabor the point, but being on the same quest as yourself I decided to follow up on this. I pulled the jobs reports days from the BLS website dating back 20 years. The period from 1996 through EOY 2000 was great for this idea, but other than that the equity chart is mainly down. From 2001 onward I calculate that this has lost over 60% of the profit gained from 1990 up until that point. In my testing a higher close on the session preceding the jobs report didn’t impact the win rate or profit factor.

          Using $INX in TradeStation, I calculate 120 trades (237 trades without higher close criteria), 55% winners, 1.26 PF for 1990-2010 with the max draw down of about 2x the net profit. Using SPY or continuous ES contract shows similar results. YM, NQ and TF (Russell 2k) are all significant losers. I would be happy to share the details of my code and testing if you would like.

          I would like to find an edge here, but to me the draw down is just too much for me to feel comfortable. As always, I could be mistaken and welcome any of your experience and insight.

          Thanks,
          JP

          • JP,

            1) the SPY has to be adjusted for dividend and cash payments,
            2) ES, YM, NQ and TF close on 4:15 PM, not 04:00 PM (which sometimes makes quite a difference),
            3) as stated on the posting the setup is triggered on the session immediately preceding the first Friday of the month, regularly -but not always- Job Report Friday (e.g. in July they regularly don’t match), assuming (but not being sure) that any deviation would be averaged out over the course of time.

            Please check the last 7 occurrences, the SPY should be up on the first Friday of a month when it closed higher on the previous session (even it would’ve been a session at the end of the preceding month).

            Best,
            Frank

  4. […] in a microcausm this is also responsible for the end of month/first of month effect as well. http://www.tradingtheodds.com/2010/01/how-to-make-millions-in-trading-the-spyder-seasonalities-i/ There are even strong seasonal tendencies surrounding the holidays and rising optimism prior to […]

  5. […] in a microcausm this is also responsible for the end of month/first of month effect as well. http://www.tradingtheodds.com/2010/01/how-to-make-millions-in-trading-the-spyder-seasonalities-i/ Further confirming this “optimism effect” are the strong seasonal tendencies […]

Leave a Reply

Your email address will not be published. Required fields are marked *