Wednesday, 27 November 2019

Strategy Validation with Dave Bergstrom

With the toolsets we have available to us today it’s really quite easy to create a trading strategy by just mining market data.

Build Alpha: As we’ve just heard in that opening bit of audio and also from previous podcast guests too, if you try enough combinations you can find something that appears to work purely by chance or by luck.

The challenge however is trying to identify something that could be sustainable.

Something that may persist long enough in the future for us to take advantage of, and hopefully make some money from.


Our guest for this episode, Dave Bergstrom from BuildAlpha, has spent years researching, building, testing, and implementing market making and trading strategies for a high frequency trading firm, CTAs, money managers, individual clients, and even aspiring retail traders.


In this episode Dave is going to share some of his insights into strategy development and validation, including:
  • How adjusting the ratio of in-sample/out-of-sample data can lead to creating different types of strategies
  • Variance testing – what is it and how can it be used in the strategy creation process
  • How E-ratios can be used to determine how an edge decays over time & weed out potentially poor strategies with good backtest results
  • Why volume and volatility are important factors to consider when building trading strategies
  • Loads of other ideas to test and validate the robustness of trading strategies.

Tuesday, 19 November 2019

Greatest Traders of All Time

As a quantitative trader, I could not have been more excited for the new book “The Man Who Solved the Markets” by Gregory Zuckerman which details Jim Simons incredible story.

Jim Simons averaged a 66% return over the past 30 years and a 39% return after his 5% management and 44% performance fee (pg 316 of book).


I plowed through the book and had, what I believe, are some major takeaways to share:

1. Edge is important; not the story of why it exists. 

In other words, data mining is ok.

This is something I’ve long defended since the launch of Build Alpha. You do not need a hypothesis or explanation of why a certain investing/trading edge exists if it is statistically relevant or significant.



In my opinion, it is possible we simply cannot comprehend why a pattern or edge exists because it exists in a dimension too complex for our current understanding. Therefore, we should not discard edges that we do not understand.

This is why I (and BuildAlpha) search the market for edges and let the data tell us where the edge is. Remove the human bias, false ‘truths‘ and the need to explain/justify everything with a hypothesis or reason why it is happening. Many of these patterns are ‘overlooked’ because they don’t have an explanation, but have clearly been profitable for Renaissance!

Here are a few quotes to drive home takeaway #1:

“Simons and his researchers didn’t believe in spending much time proposing and testing their own intuitive trade ideas. They let the data point them to the anomalies signaling opportunity. They also didn’t think it made sense to worry about why these phenomena existed. All that mattered was that they happened frequently enough to include in their updated trading system, and that they could be tested to ensure they weren’t statistical flukes”. (pg 109)

“Simons and his colleagues hadn’t spent too much time wondering why their growing collection of algorithms predicted prices so presciently. They were scientists and mathematicians, not analysts or economists. If certain signals produced results that were statistically significant, that was enough to include them in the trading model” (pg 150).

“I don’t know why the planets orbit the sun. That doesn’t mean I can’t predict them” – Simons (pg 151).

“More than half of the trading signals Simons’s team was discovering were non-intuitive, or those they couldn’t fully understand. Most quant firms ignore signals if they can’t develop a reasonable hypothesis to explain them, but Simons and his colleagues never liked spending too much time searching for the causes of market phenomena. If their signals met various measures of statistical strength, they were comfortable wagering on them.” (pg 204).

“Volume divided by price change three days earlier, yes, we’d include that” – Simons (pg 204).

2. Everyone struggles with discipline and following their system. Even the Greatest of All Time (G.O.A.T)

Discipline is key and the ability to consistently follow your system(s) can be the difference between winning and losing. We all believe discipline becomes easier if you have more reliable edges or have grown your account quite a bit, but Jim Simons would probably argue that is simply not true!

BuildAlphaIn Dec 2018, Simons (worth approx. $23B at the time) called his advisor and wanted to override his systems (pg 308). The systems that have created the most incredible track record in history.

In the “Quant Quake” of 2007, Simons overrode his systems before the eventual rebound. One employee was quoted as saying it cost the firm money (pg 260). Moral of the story.. follow your system and trust your research! Everyone struggles with this, but we must.

Note: Majority of his career Simons was actually the one advocating to NOT override the systems and may be a large part of his success. These were just two small examples.

“Trust the model. We have to let it ride; we can’t panic” – Simons (pg 216

3. Surround yourself with a great team

This one should be obvious, but no one becomes the G.O.A.T alone. Brady has Belichick, Jordan had Pippen, Kobe had Shaq, Ruth had Gehrig, etc.

A large portion of the book chronicles how Jim sought out help from brilliant individuals, hiring them away from prestigious positions (science, tech and academia) by offering to double their salary. I won’t go over every individual, but a lot of chapters in this book are dedicated to the spectacular individuals that helped create the incredible returns which give Jim Simons the G.O.A.T title.

He recruited great talent to his team. Surround yourself with those that are experts in things you are not or inspire you to push past your limits.

Incorporate different approaches to your own similar to how Simons did. Trading is a lonely business at times.. you don’t need a hedge fund to build your own team.

4. Build strategies using different data.

Sure price and volume are great but the book mentioned other areas of alternative data Renaissance found useful.

Here are some simple ideas the book mentioned:

– sentiment
– correlations and relative moves
– number of times a stock’s ticker appears in major publications (regardless of sentiment)

Additionally, here is a previous See It Market blog I did using Commitment of Traders report to generate a trading signal: https://www.seeitmarket.com/how-to-improve-market-returns-using-alternative-data-17806/

5. Edge doesn’t have to be big.

Renaissance searched for “overlooked” edges and joked about a 50.75% win rate while utilizing the law of large numbers to win in the long-run.

Often times we get caught up searching for the holy grail or the perfect entry/exit for our trading or strategy development. But even with all these PhDs, RenTech was excelling trading a nearly 50% winning system to generate such astronomical returns. Much more can be gained by combining and adding unique smaller edges together than wasting time hunting for the perfect holy grail strategy!

 “We’re right 50.75 percent of the time… but we’re 100 percent right 50.75 percent of the time. You can make billions that way” (pg 272)

Bonus:

Build Alpha: Money isn’t the be all end all. He’s had tremendous tragedy in his personal life. Remember to enjoy LIFE while on the financial quest we are all on! The market isn’t going anywhere.I enjoyed the book and hope you do/did as well.

Sunday, 17 November 2019

Tracking Sigma Scores Of Price Changes For Regime Shiftsh

Measuring price moves is the name of the game.

However, measuring price moves given recent context can add additional benefits to your trading performance.

Build Alpha: At times, the market can get very quiet which can make a 1% price drop feel like a 10% price drop (think summer trading). At other times, a 10% price drop can feel like a 1% price drop (think 2008). This is a driving force in why I often prefer to view price moves in standardized form and expressed as “sigma” moves. A sigma score basically tells you how many standard deviations can fit between the mean and the current price move. To calculate sigma you simply subtract the mean from the underlying value and then divide that difference by the standard deviation.

Sigma = (X – Mean) / Stdev

In our case we will have X = the natural log of the one bar price change or log(close[t] / close[t-1])
The astute reader would question the lookback period used to calculate the mean and standard deviation of X used in our sigma calculation, and that brings us to a very unique indicator I often view.

I personally like to track the rolling window of one month (22 trading days) and one year (252 trading days). I then like to compare the one month sigma score of price changes to the one year sigma score of price changes. This gives an indication of how current volatility compares to more long-term volatility.


Below is a plot of the sigma score calculated on a rolling monthly basis and on a rolling yearly basis. You can see the monthly sigma scores stay “bounded” between -4 and +4 whereas the yearly sigma scores vary a tad more.


All of this is great, but how can we use this and why am I telling you about this…

Build AlphaWell when the monthly sigma minus the yearly sigma difference becomes greater than 1.5 we start to identify some key trading periods or moments in the S&P 500. For example, the instances where the difference between the monthly and yearly sigma reach 1.5 or more include the peak in 2007, the bottom in 2008, the flash crash in 2010, the European Debt Crisis of 2011, and the ETF meltdown of 2015.

Yesterday, May 17, was a significant down move but we are not near “significant” moments in history (yet) as the yearly sigma score stays subdued. This is definitely something to keep an eye on if this volatility persists throughout the summer.

Over at Build Alpha, I produce software that automatically creates systematic trading and investing strategies, allows the validation and testing of each strategy, and generates exportable and executable code for each strategy – no programming necessary.

Sunday, 10 November 2019

Providing efficiency in Trading

Trading is a phenomenon which is not a cake walk. One cannot become an expert in this field unless one has gained some experience and knowledge in it. You need to seek guidance in the initial stages of trading. For taking help you can either reach out to some expert in this field or find advanced trading software that comes with training. As the trading in securities markets is based on interpretation of the market forces which in a lot of sense depends upon the market forces so the human interpretation can lack at times in analyzing certain aspects of it.

This is where software like Build Alpha comes into play and can help regulate your operation. This is a software which helps us in making strategies, testing our ideas, and building our confidence to run automated strategies in real-time. This is one of the best trading softwares to date and serves the needs of traders. Anticipating the trend or mean reversion becomes much easier if you do it through build alpha software solution. The best part… no programming is needed!



Trading software is a computer program designed to help traders and investors to trade and analyze stocks, commodities, futures and various currencies in forex market. It also gives you access to the latest news that affects the price movements of stocks and their trends. There are different types of trading softwares for both beginners and experts. BuildAlpha is one such software which is helping the traders to achieve the success in trading regardless of their skill level. It allows you to stress test each and every strategy.

View in and out of sample test results with ease, has data included, and generates actionable trading code all from one command center. There are fund managers who study short and long term charts with the help of these softwares. There are softwares available to know technical and fundamental aspects of the stocks. Many stock brokers provide softwares to their clients to trade on their own. The main features of the software are order placement and proper analysis of stocks. These softwares are built by third party and licensed to online brokers.

The software gets new updates every week which makes it more unique and worthy than other softwares which are alike in the trading market. The new peculiarities that are coming out every now and then are amazing it seems like trading has become easy with the aid of this software. You will experience the difference after using the Build Alpha software. You will become a person who will start analyzing what is important. Get efficiency in trading markets with the expert analyst feature of this software, which helps you in analyzing trends more specifically.