Monday, October 25, 2010

Golden Cross Test

I know it has been done before, but I wanted to test the "Golden Cross" theory for myself.   A golden cross is when the 50-day moving average for an index crosses above the 200-day moving average.  Market strategists and financial media pundits really tout this event, saying that is very bullish for the future.  I think it is just a good way to get a retail investor's attention, say 'golden cross' or 'death cross' and people tend to get excited because they don't know any better.  On October 22, 2010, the S&P 500 hit its Golden Cross. Below is a test I ran to buy the S&P500 index when it hits a Golden Cross, then sell 'n' days later. Data is from Yahoo Finance and the results are frictionless.  Go here if you can't see the spreadsheet.

Results:
 
Buy S&P 500 when 50 day MA crosses above 200 day MA; sell 'n' days later. $100,000 per trade; 1951 to present.
Period
Exit n
Days
Net Profit
# Trades
# of
 winners
% of
 Winners
Max. Sys
 % DD
Avg %
P/L
W. Avg.
 Profit
L. Avg.
 Loss
Profit
 Factor
Payoff
 Ratio
1 week
5
         6,191.24
30
15
       50.00
-7.10
0.21
      1,509.92
  (1,097.17)
1.38
1.38
1 month
20
      34,537.47
30
19
       63.33
-14.12
1.15
      3,385.79
  (2,708.41)
2.16
1.25
3 months
85
   161,879.05
30
24
       80.00
-12.78
5.4
      8,766.50
  (8,086.16)
4.34
1.08
6 months
125
   190,057.46
30
23
       76.67
-17.83
6.34
   10,451.34
  (7,189.04)
4.78
1.45
9 months
185
   295,523.68
29
23
       79.31
-16.59
10.19
   15,116.42
  (8,692.33)
6.67
1.74
1 year
250
   298,919.42
25
20
       80.00
-16.12
11.96
   16,737.79
  (7,167.29)
9.34
2.34


The results are pretty good, not hugely robust, but decent nonetheless. It appears that the Golden Cross is best used a longer term indicator. Performance over 3 to 12 months is much better then 1 week or month.  It seems that if you buy the Golden Cross, then sell 3 months later, you'll win 4 out of 5 times and suffer mild drawdowns. This strategy completely avoided the destruction of 2007-2008 and was 100% successful from 1979 forward. So it seems this Golden Cross theory does have validity.

Now if you buy the S&P 500 when the Golden Cross hits, then sell when the Death Cross [200 day MA crosses below 50 day MA]:

 
Buy S&P 500 at Golden Cross, sell 'Death Cross'. $100,000 per trade; 1951 to present.
Net Profit
# Trades
# of
 winners
% of
 Winners
Max. Sys
 % DD
Avg %
P/L
W. Avg.
 Profit
L. Avg.
 Loss
Profit
 Factor
Payoff
 Ratio
Sharpe
 Ratio
   494,928.99
30
24
80
-12.88
16.5
   22,344.02
  (6,887.93)
12.98
3.24
0.33

It is not shown here but the average holding period is 326 bars, so that is basically 15 months. That is a pretty long time for a strategy to come to fruition. There are many ways to make money in the meantime, but this one is still successful if it is all you've got.

Conclusion:
In this simplistic test of the Golden Cross theory, we can see that it is indeed bullish. It functions as long term, not a short term, indicator. If a buy and hold investor wants to time his purchases, then following the Golden Cross could be a good starting point.
What about shorting when the Death Cross hits? I'll deal with short in a future post. The results may not be what you expect.
Footnotes:
Test was run in Amibroker. The portfolio buys exactly $100K each time, regardless of the portfolio size. Bottom line, it is not compounding. This reduced gain metrics and understated the drawdown calculations as compared to a fully compunding portfolio. Also, as mentioned above, the results are frictionless. There is no slippage or commissions; I just wanted to test a theory.

1 comments:

xtremepicks.com said...

Definitely seen a lot of bullish trends after a golden cross when we feature them on http://www.xtremepicks.com - CAVR is another alert just forming this golden cross - time will tell but looks like its ready for a bullish run.

Post a Comment