Algorithmic Trading Strategy Review: Short Term Trading Strategy Offers Excellent Alpha Over the S&P 500 During the Month of July.

The article was written by David Shabotinsky, a Financial Analyst at I Know First, and enrolled at an undergraduate Finance program at the Interdisciplinary Center, Herzliya.

Algorithmic Trading Strategy Review: 

I Know First, Ltd. is a financial technology company that provides daily investment forecasts based on an advanced, adaptable, self-learning algorithm. The company’s algorithm predicts over 3,000 securities (and growing) with capabilities to discover patterns in large sets of historical stock market data.

Since the beginning of 2016, a new short-term strategy had been implemented, which focuses on optimization of the forecasts for up to one week. This article will review the highlights, specifically the performance of the model as well as of the daily stock selection processes utilizing the short-term signals for the most recent month of July in more detail. The full description and report on the respective strategies analyzing their statistics since the begin of the year can be found here.

Each trading strategy uses a different ranking function based on the short-term signals. Throughout the year the daily rebalanced trading strategies had successfully selected the “best” of the S&P 500, offering a high alpha of up to 35.79%. The overall Sharpe Ratio, which is a financial ratio of the risk premium (above a risk-free rate) over the volatility of investment, was high as well, as shown below. The high Sharpe Ratio indicates high return against the risk taken on the investment. Therefore, the active strategies implemented by I Know First, clearly show not only a higher a return than a passive investment into the S&P 500 but also a lower risk per return.

The strategies clearly outperform the market, as represented by the S&P 500. Overall, the trading for the month of July had up to a 4.86% alpha above the market.

To understand the outperformance in greater detail, a closer look is taken into account into the raw one-day signal statistics. Initially a predictability filter is applied to focus on the more predictable stocks within the S&P 500 universe. The daily threshold changes daily depending on the predictability level of all the stocks in the system on a given day. Usually, approximately half of the S&P 500 stocks pass the filter, based on the highest predictability level. To further analyze the performance of the signals they are divided into percentiles.

The Table A below depicts the daily trades returns based on the strongest (top) 10% and 5% of the signals, that strongly outperform the average daily return one would receive by taking equally weighted long positions in all the S&P 500 stocks. This is the bottom line of the algorithmic performance.

Key Results- July 2016 (pilot project): Table A: Predictability filtered signals 01.07.2016-29.07.2016

Table B depicts the actual strategy at work. For example, the most intuitive selection process IKF_Top20_pure results in 20 stock picks daily, with the strongest predictability filtered signals found in the top 10% and 5% percentiles of Table A. They are then traded and rebalanced on a daily basis, yielding 7.11% return for July 2016 with 59.25% of winning trades among the 400 selected. Other strategies apply more complex rules using signal and price dynamics, which remained unchanged since the beginning of 2016 and thus do not present over-fitting results here. Remarkable for those is the high ratio of the average winning to average losing trade, reaching up to 1.54 in case of IKF_Top20_acc_cons_filter which yields 0.45% per trade on average.

Table B: strategies performance 01.07.2016- 29.07.2016 (=pilot project). Overall return IKF between +5.13 and +7.44 (8.42%) vs. S&P 500 +3.56%*avg real return- the average daily close-to-close return of the S&P 500 in July 2016

As recently mentioned in the article Short-term Trading: Summary of Realistic Backtests based on Daily Stock Selection, the strategies are not fully exposed to the overall market, with beta being below 0.5. Additionally, all of them generate a great positive alpha and keep the Sharpe ratios high. Hence the I Know First short-term signal based strategies offer a great opportunity to generate additional high risk-adjusted returns, while further diversifying the portfolio.

I Know First Research

[email protected]

People who read this article had read as well:

  1. An Analysis of the I Know First Signal and Predictability Indicators
  2. Stock Forecast: Daily Stock Selection Based on a Self-Learning Algorithm 
  3. Daily Stock Trading Strategy Returns 42.59% in H1 2016
  4. Short-term Trading: Summary of Realistic Backtests based on Daily Stock Selection
  5. I Know First Stock Forecast Algorithm