An Intro to Quant Models: Part 3

This article forms part of a series by CEO Sam Barry on Quant Models. To read the first two parts, follow the links below:

Last time round, we spoke about the models we use in creating systems and strategies, so this time I want to  start to touch on the types of quant trading strategies and, just in case there are some budding strategy designers out there, we would love to hear from you.

People describe strategies in numerous forms, but actually I suspect there are only a few real categories and styles.

Firstly you have trending strategies and counter trend strategies.

Essentially, there are two real concepts in trading, those that attempt to capture the trend and those that try to capture deviations from the trend.

From a quant’s point of view, all trading strategies are based on inputs. Whether it is from technical indicators or fundamental data releases is somewhat irrelevant, so I tend not to think in terms of Fundamental vs Technical beyond the reasoning for why the system works.

Beyond Trend and Counter Trend, we have Static and Adaptive systems.

In Static systems, you set the parameters and that is it; you either have it on or off. An Adaptive system typically will do one of two things, it will either change its own parameters given some inputs, or it will progressively learn.

Machine learning is often thought of as some complex process but it really isn’t, all it is, is ongoing optimisation.

Then finally, we have hedging and non-hedging systems. Essentially, these are systems that can take opposite trades at the same time or not. We call them hedging here at Littlefish FX, but to be honest it’s because we don’t have a better name for them.

You can then combine these categories to define your system. For example, take a Trend Trading Adaptive Hedging strategy; essentially this is a system that uses a set of inputs to trade a trend, it learns as it progresses and can take a long and short in the same underlying at the same time.

The key to all of these strategies though, is actual simplicity.

I would say 95% of the strategies used in most quant models are really simple and there are good reasons for it.

Firstly with too many parameters there is a real risk of curve-fitting. This is possibly the biggest danger of any quant system in my opinion; that your system is so well defined by past data that doesn’t work in the future.

The second reason with simplicity is the amount of computing power required to do a lot of the calculations involved with modelling. If you start layering parameters, the computing power / time grows exponentially.

The third reason is that, actually, a lot of the simple strategies work really well.

Take moving average crosses; nothing could be simpler, but if applied correctly and layered within a model using good inputs, these can be extremely effective trading tools for trading trends.

I always get annoyed by these ‘educators’ that suggest pro traders use 200 inputs and indicators to trade the markets. It’s utter nonsense: you just could process it either as a human or a computer not unless you are sitting in IBM or Apple burning through servers.

The reality always with trading is that the simple systems still get you well defined probabilities of accuracy, and with well-defined probabilities you can set good risk and money management rules to make the systems profitable. If you can then execute in various conditions through this, then that’s all you need.

So if you want to build a quant system for us, consider these key four questions:

  1. Is it designed to trade trends or counter trends? Trends happen more frequently but can be difficult to judge.
  2. Are the parameters static, or will the system learn, and if so how?
  3. Is it designed to take multiple positions in one underlying or one position?
  4. Can you explain why it would work? Note: this is the most critical, if you can’t explain why it should work then the concept is flawed.

Stay tuned for Part 4, on building trading systems and managing risk. Meanwhile you can catch up with my previous articles on the subject of quant models here:

Got a question for Sam? Tweet him directly at @LFXSam or contact us here

Introducing our new Forex Trading Course: with indicators, trading plans & strategies!

Get your hands on all the trading plans, indicators and strategies required to gain a comprehensive understanding of the Order Flow trading techniques successfully used by the traders here at Littlefish FX. Find out more here.