Blog - Investing Notes

February 7, 2006 – Collective2: A Marketplace of Trading Systems

Declan Fallon of Fallond Stock Picks Inc. alerted us to a third party verification system, Collective2, that he uses to present a model portfolio. Matthew Klein operates Collective2 as a marketplace for trading systems, offering shoppers a selection of "over 1,200 trading systems for stocks, options, futures, and forex" and suppliers a "place to market and sell trading advice" that offloads "the drudgery of subscriptions, emails, credit card payments, and running computer servers." Does the aggregate trading record at Collective2 support the Efficient Market Hypothesis? Using statistics available there for 143 active stock trading systems (as of 2/3/06), we find that:

Collective2 provides the following statistics for each trading system offered:

• The length of time, in weeks, that Collective2 has been tracking the system (ranging within sample from 1 week to 121 weeks);
• Whether it is a daytrading (rapid-fire) or swingtrading (weeks per trade) system;
• The percentage of trades that are profitable according to real-time price feeds (ranging within sample from 0% to 100%);
• Cumulative dollars won divided by the cumulative dollars lost (ranging within sample from 0 to 65.8);
• Sharpe ratio (ranging within sample from -16.0 to 2.94); and,
• Raw percentage return on account equity, non-annualized (ranging within sample from -91% to 655%).

Accounting for the likelihood that outperforming systems persist and underperforming systems quickly disappear, we use an equal-weighted average return per week across all trading systems as the principal measure for aggregate Collective2 marketplace performance. We calculate the average return per week for each trading system by dividing the raw percentage return by the number of weeks tracked. The following chart shows the distribution of the 143 active stock trading systems by average return per week. The overall average of averages is close to zero.

The next chart shows the overall average of average returns per week for all 143 trading systems and for several groups of trading systems within the sample. Daytrading underperforms swingtrading. Aged systems greatly outperform immature systems (probably due to extreme survivorship bias, as noted above). Benchmarking these groups of trading systems is difficult because of the different trading timeframes involved. The average weekly return for the S&P 500 index during 2005 (total return for the year divided by 52) is 0.06%. The average weekly return for all active trading systems (presumably before transaction fees) is therefore about the same as that for the S&P 500 index during 2005.

The next chart shows the average Sharpe ratio for all 143 trading systems and for several groups of trading systems within the sample. On a risk-adjusted basis, swingtrading greatly outperforms daytrading. Aged systems again outperform immature systems, very likely due to survivorship bias. Again, benchmarking these groups of trading systems is difficult because of the different trading timeframes involved. Using weekly data, the Sharpe ratio for the S&P 500 index during 2005 is approximately zero.

The final chart shows the percentage of winning trades for all 143 trading systems and for several groups of trading systems within the sample. All results exceed 50%, with the only notable difference that swingtrading again outperforms daytrading. As a rough benchmark, during 2005, the S&P 500 index has positive returns for 30 out of 52 weeks (58%).

In summary, the aggregate performance of the stock trading systems active on Collective2 as of 2/3/06 is not inconsistent with the Efficient Market Hypothesis.

For summaries of related research, see our blog entries of:

10/31/05 finding that some skillful individual investors do exploit market inefficiencies to earn abnormal profits;

10/17/05 providing a hedge fund benchmarking analysis;

6/24/05 concluding that the active management component of broadly diversified mutual funds is generally expensive and ineffective;

2/4/05 finding that the actual aggregate (timing) experience of equity investors is inferior to passive buy-and-hold stock market returns;

2/2/05 for a wide-ranging (and critical) review of the aggregate performance of mutual fund managers by Burton Malkiel;

10/29/04 on the tendency of individual investors to underperform by overtrading; and,

10/22/04 finding that over 80% of day traders lose money after transaction costs.