The opinions expressed in these forums do not represent those of C2, and any discussion of profit/loss is not indicative of future performance or success. There is a substantial risk of loss in trading. You should therefore carefully consider whether such trading is suitable for you in light of your financial condition. You should read, understand, and consider the Risk Disclosure Statement that is provided by your broker before you consider trading. Most people who trade lose money.

New Trading Strategy: ETF Sentiment Magic US

Hello, I would like to introduce a new trading strategy that we put live here on Collective2 since last week. This is based on a quantitative-based algorithm. It uses proprietary alternative data to “predict” the market sentiment for the next month. We believe the source of alternative data that we are using will not decay and cannot be influenced by parties. It is a relatively “boring” trading strategy but one with succesful results.

We use 3 times leveraged ETF’s that are based on the Nasdaq index, S&P500 index, Gold and US Treasuries. All ETF’s are listed on ARCA.

The algorithm distinguishes between 6 different market regimes being Super Bull, Bull, Bear and 3 more stable market regimes. Based on the identified market regime, it will invest in a specific combination of ETF’s / indextrackers. So in markets within turmoil it will move to a Bear market position or move to US Treasuries or Gold.

Backtests have shown the following results over a 15-year period:

  • We target an average return of 80-85% per annum
  • The algorithm is performing the best in trending markets, either upwards or downwards. In more stable markets, the performance is also good, but it does have more difficulty in accurately predicting the direction of the market then (which is logical, hedge funds will also show worse results in those years and a Buy-and-Hold in that market also will not show superb results).
  • Over the full 15 year period, the algorithm is showing profitability in every single year. This is quite unique for a strategy.
  • The max drawdown over the full 15 year period is comparable to the drawdown you would have had with a buy-and-hold strategy on the (unleveraged) S&P500 index over this period.
  • The algorithm is designed to trade daily, It will adjust the predicted market sentiment once per month, unless there is a reason to do it earlier (when all hell breaks loose, it will switch earlier). Daily rebalancing is done to increase the returns and lower risks.
  • The algorithm will only use a short position in the Bear market regime. It will short the S&P500 3-times leverage ETF in that regime. All other market regimes only use Long positions.
  • Our professional team is monitoring the performance of all of our algorithms daily.

This strategy has been designed to be suitable for investment amounts from 5K upwards to very big volumes. The ETF’s we are using have a very big liquidity so we hardly have any impact on the market with our volumes.

Below is the URL to this trading strategy:

For people that are interested we do have a PDF available with the backtest results from 2014 to July 2019.

Sorry, I meant the backtesting results from 2004 to July 2019 so including the 2008 recession (where it actually did very well). Only in 2014 the algorithm had difficulty with achieving a high performance.

You should not use leverage, a good strategy includes also low volatility

Hello Jose, technically we do not use any leverage since the algorithm never invests more than 100% of the cash available. However, we do use ETF’s which have a leverage of 3.0 build-in. That is where the leverage comes from.

If you are interested, we also have a version of this same algorithm which uses the unleveraged versions of these ETF’s. This algorithm therefore has a leverage of 1.0 and the average annual return there is 21-22% per year. Sharpe ratio around 1,9 - 2.0.

If you are interested, we can publish this one as well. Then people can choose whether they want to go for the unleveraged (less risky) version (with lower returns) or the 3.0-leveraged version (with higher returns). Both are very stable anyway…

Hello Jose,

Today we also put Public the unleveraged version of this same strategy. See below.

This strategy invests in the US-listed unleveraged versions of these ETF’s. We have also been able to backtest this over 19 years, from 2000 up until August 2019 with an average annual CAGR over this period of 14,5%. The 3 times leveraged versions of these ETF’s only exist as of 2009 but the unleveraged versions exist for a much longer period. The maximum drawdown over this whole period is also limited and in most years it stays below the 10%. In 2008 it was around 15% and the maximum drawdown in 2011 was 24,5%. All other years below 10%. With a Buy-and-Hold investment in the S&P500 the maximum drawdown over this same period would have been much higher. 2008 and 2009 actually were some of the most succesfull years because the sentiment prediction moved to the right sentiments (Bear sentiment etcetera).

This strategy is suitable for longer-term investors and it only uses long positions. So could be useful for pension related investments, 401K’s etcetera. If you are interested, just ask for the backtesting results.

To avoid issues of overfitting / overtraining, we have used all the necessary precautions. So we build and tested it on a different period, after which we validated it (after development) on a completely different period. We also run this strategy with similar results for 6 months in paper trading. And we already finalized this strategy in August 2018 and it also shows very good results after this period, so we have all the faith that this strategy is not overfitted or overtrained. We are trading the EU version of this strategy ourselves. We are based in Europe so unfortunately we can only trade in the EU-listed versions of these ETF’s.

1 Like
Back to C2 Platform