A short only gold strategy: Breaking the Momentum

Gold has been used throughout history as money and has been a relative standard for currency equivalents specific to economic regions or countries.  It is the most popular hedging instrument againts economy.  When bad news starts coming to market, gold price starts increasing.

Here is a chart og gold price from Jan 2000 to  August 2010.



Trend is definitely upword and few investors might dare to go short in gold. However gold has been volatile for last 3-4 years.  You can see it from chart from begining of 2006 to end.

My idea is to make money when gold price falls.  Only way I can do it by going short for a minimum period (Say 1 day).

Strategy:
If gold price is going upward continuously for a  good number of days, there is a possibility that the trend will break on the next day.
I  used  data from 2006 to 2010 and put a threshold for each up-move to make sure the trend is strictly increasing.
Looking at  pevious X days return and going short for a day, I got good return

Average return: 0.43%

Total number of trades: 21

No of possitive retun: 17

Standard deviation: 0.87%

Win rate: 81%

This is the chart for wealth



I could not get more signals, however you can look back less number of days and adjust threshold to lower level to get more  number of entry signals. 
[I could get only 21 entry signal from  2006 to 2010 and I have not used any transaction cost. Strategy may not work for future time period. ]

Simple Strategy for Indian Market : 1

Long time back, I was through an interview process. The interviewer gave me a pair of equities to formulate a trading strategy.
I worked on it  for a week and got a very simple strategy which should work on real market. However, by that time I had completely made my mind to stick to my organization. Yesterday, I made my mind to do further analysis of the strategy. To my surprise, it worked better than my expectation on my OpenOffice Calc. (I hate MS Excel as it does not come free with Windows OS)

I downloaded data for a pair of instruments which are very much correlated. Almost identical (Hint: One Sector Index future-A and another may be ETF-B).

Strategy:
A and B should move together. If A and B deviates,  there is an opportunity.
This is a simple statistical arbitrage strategy.

Ra(t)= Return of A for last t days
Rb(t)= Return of B for last t days

Ra(T)= Return of A for next T days
Rb(T)= Return of B for next T days

th= Threshold (1,2 3 percentage)

If |Ra(t)-Rb(t)| exceeds 'th', we have signal to take position.
One scenario:
Long in A  and short in B if sign(Ra(t)-Rb(t)) is -ve.
Result for last 4 years is like below: (With 0.5% transaction rate)

Mean 1.19% Win % 82.20% Max 7.04% Min -2.36% Stdev 1.69%

Wealth graph over 4 years(Starting from 100 Rs)

Wealth (Only for year 2010)

However,  in market similar result may not be achievable. ETF has tracking error which may give false signal. There is also a limitation to investment amount as ETF is illiquid in Indian market.