Order Book Visualization

Charting tools suffer from a lack of transparency. Bookmap's novel solution to Limit Order Book Visualization zooms in on actual supply and demand.

Tsachi Galanos of Bookmap describes the firm’s novel solution to Limit Order Book Visualization and analysis

Roughly two years ago, we shifted our company’s focus from proprietary trading activity, dealing mainly with HFT algorithms, to creating an Order Book Visualization platform. We named our platform Bookmap (stands for order BOOK and heat MAP). The journey of the past 2 years taught us a lot about the importance of order book analysis and visualization to both traders and quants. I would like to share some of the insight we’ve gained in the process, will detail about Bookmap solution as well as few thoughts about the future evolution of visualization.

The problem with current charting tools

There is an inherent lack of transparency in current charting tools. They provide filtered or aggregated perspective of the market and therefore do not allow traders and quants to get real insight into actual supply and demand, leaving them shortsighted to the activity in the market.

Why look at the order book?

The limit order book is the place where all buyers and sellers meet and where the auction takes place. It is basically the entire supply and demand. When we look at the order book, we see all the decisions of all the market participants and the strategies they employ. In contrast to looking at the volume traded, which shows what already happened, the order book provides an insight into the intention of the traders. This information can be used for forecasting short term price action.

Here are few examples of the type of information that can be gained by following and analyzing the order book:

  • Find price levels with Certain price levels attract large numbers of orders. These clusters can be identified only when looking at the depth of market. In many cases, these levels will act as support and resistance, depending on whether they are above or below the current price.
  • Watch the spread which can provide clues about what might happen in the immediate future.
  • Identify intraday shifts in momentum, for example a shift from strong buying to strong selling.
  • Price changes due to order book activity. Study the relationship between market price movements and the order

When using the depth of market we can also observe strategies in action, for example:

  • Price triggered strategies. Strategies that automatically change orders positions based on price change.
  • Validate breaks of key technical analysis levels and better assess if the breakout is real or false.
  • Identify when big players step
  • Identify retail

There are varied ways to use the order book. Scalpers may use the information from the  order book to decide whether to go long or short; Swing traders or technical analysis driven traders may use it to confirm their macro buy or sell decisions.

The challenge when dealing with limit order book

Most quants and traders who use order book analysis have to deal with the following challenges:

  1. HFT strategies are placing limit orders on many price levels just to be first in queue. Usually the orders are placed when price changes. Most of these orders are later canceled when price gets
  2. Not all the limit orders that are sent to the market indicates an intention to trade. There are traders who send orders to manipulate price and create fake liquidity, such as spoofing or quote
  3. In some cases not all the orders are visible in the order book. There are types of orders in various exchanges that are

Some of these challenges can be dealt with and some are more complex. More on this below.

Why visualization is important?

Visualization plays important role in understanding what is really happening and in taking better and more educated decisions. Looking at data in a pictorial or graphical format enables us to grasp difficult concepts with greater ease or identify patterns that are otherwise unobservable. It also helps us to ask questions we did not ask before.

As a result of technological advancement, visualization of big data is now possible. With current CPUs & GPUs visualization that in the past could have only be done offline, can now be achieved in real time. This in turn enables faster more relevant and beneficial decision making. When dealing with market data, I see the use of visualization in the following scenarios:

  1. In early stages when cleaning and ‘checking’ the market data. In many cases wrong decisions are taken because the market data is not accurate, therefore, visualization should be considered at that
  2. When developing a new strategy or refining an existing one. In other words, when ‘teaching the computer’ what to do. Looking at visualized market data can also be beneficial in getting new
  3. When trading live and facing the need to take immediate
  4. When monitoring / watching the market in real time and take decisions like: whether to trade or not in certain market conditions or even see that your strategy is ‘not aligned’ and needs to be

In all of the above scenarios, having the right tools is vital.

Looking at the microstructure

Each macro event is a combination of micro events. In many cases, if you manage to understand the microstructure, than you will be able to better understand the macro. The advantage to study the microstructure is that there are less events, and therefore it is easier to interpret the activity and intentions of market participants. The best way to learn the microstructure is by using visualization tools and zooming  to the basic blocks of the market data.

How we visualize the order book

We had to deal with this question when we developed HFT strategies. We wanted to better understand other types of market participants and also see what happens in the market when we send our orders. We decided to visualize the order book using a heatmap, which is updated very frequently (video­like 25­40 FPS). The heat map records and visualises every change in the order book by displaying it on a scale of gray shades. The brighter shades mark price levels with larger number of resting paper while darker shades mark areas of lower liquidity.

The heatmap gave us a clear view of how the entire limit order book and traded volume evolve over time enabling us to get faster and deeper insights into market dynamics. Let me explain it further. Regular charts, such as bar chart, are two dimensional, (price and time). When you use a heatmap you add another dimension so in this case it also let you see the historical sizes at each price/time. In addition, by updating the chart very frequently (40 updates / second) you get a video which lets you view also the frequency of the changes, giving you a ‘feel’ of market accelerations.

Visualization lets you see patterns that cannot be seen or understood without it

Below are a few questions that can be investigated with a visualization platform like Bookmap:

  1. How did the size at each price level change over time?
  2. What happened to a certain level when price moved toward it?
  3. Are there additional strong levels below or above that level?
  4. What was the volume traded around these levels?
  5. What is the activity on the other side of the book? Are there areas where the order book is not symmetric?

Example: Price Bounced Back

In the above image there is a significant concentration of limit sell orders at several neighboring ask levels

Typically in real­time we wish to examine the hypothesis that price will bounce back (at least for a while) if it reaches such level. Here are some factors that can vote in favor of that hypothesis:

  1. When the price approaches that level, the amount of sellers
    1. remains the same or
    2. becomes even larger (in this case the price may bounce even before trades occur at that level).
  2. When trades start occurring at that level
    1. more sellers join and / or
    2. we observe hidden sell orders being executed there against market buy orders

Example:­ Estimating the amount of liquidity that belongs to HFT participants

Using heatmap it is easy to notice the presence of HFT agents, and to estimate their participation rate

Example:­ Zoom in to milliseconds

A single limit sell order at size of about 180 contracts was partially executed and the rest remained unfilled

Example: Significant change in the order book

The above image illustrates instantaneous cancellation of buy orders and addition of sell orders of significant amount. This was followed by price drop. Based on what we observed in zoom in, it is likely that those orders belong to the same trader

Example: Aggregated order book view

There are also other ways to look at market data, for example via aggregated view. This let’s you see in many cases how price is affected by the liquidity

Bookmap custom solution

Bookmap is currently offered mainly to traders. Recently we launched a custom service for quants and have developed an API access to the visualization layer making it available by other markets such as surveillance, monitoring, TCA & Best execution reporting.

In addition to the full functionalities that Bookmap offers today to traders, the quant solution includes additional important benefits such as:

  • An API that enables to connect with your own data and visualize both the simulation and live This enables to track the full evolution of your orders displayed on the chart.
  • Unlimited abilities to manipulate the chart, including zooming into nanoseconds timestamps.
  • Display your own order queue
  • Inspect the behavior of your trading algo during high volatility, significant position and P&L changes, high volume rate
  • Custom indicators on Bookmap chart and on a separate pane, user defined line

Below is a diagram that illustrates the solution

Where do I see the future?

  1. More transparency ­ both organizations and individuals will keep demanding wider and more accurate data. The more data you have (e.g. from multiple exchanges) and the more details it contains, the more educated decisions you can take. As we look at the market, we already observe some of these trends taking place. A good example the upcoming launch of CME Market by Order data, which provides individual queue position and its size. This information eliminates the need for traders to calculate queue position and lets them take a more educated
  2. More analytics & visualization ­ As technology advances, more data is collected, transmitted in real­time or by demand (due to faster internet), analyzed and visualized (due to better GPU) by ordinary computers. Same as with other industries, the financial sector will require better analytics and visualization applications that will be used not only for offline research, but also in real time, with the aim of gaining quicker and better decisions.
  3. Interactive visualization, flexibility & modularity of visualization tools ­ visualization software should become more neutral to the data source and be able to display data from a wide range of sources in a meaningful way. In addition, together with the growth of data, there will be more flexible ways to engage with it. As an example, imagine that you could use your own data, select your parameters and build a video like you build a chart in Excel, augment this video with your own indicators and decide if you want to watch it offline or in real These analytics will be used not only by algo developers or quants seeking to understand their systems better, but also by traders / non ­quants carry out market analysis.
  4. Data analysis automation. More data also means more data dimensions (different phenomena types, data anomalies, different instruments, different time scales, etc.). Most of the analytic is done today in 2 dimensions but will also be available in 3 or 4 dimensions, increasing users’ insights and competitive edge. However, human perception is limited to a number of dimensions (e.g. 3D visual, audio, etc.), therefore it is logical to perform preliminary automatic data analysis and generate alerts and visualizations of most meaningful parts of the
  5. Virtual reality technology can be used to display more information. For instance, a huge trading room with dozens of monitors can be displayed virtually without the need to invest in expensive

For More information on Bookmap visit Bookmap Created by VeloxPro

Logged-in members can download the article by clicking the link under all the “Related Posts” below. If there isn’t a link then you aren’t logged in! To log in or register visit here.

Related Posts

Rigorous Optimisation of Intraday Trading: Wilmott... The progressive availability of automated access to exchanges and the continuously increasing capabilities of electronics (capture, storage and proces...
Volatility: Time and Black–Scholes–Merton The formalism of Black–Scholes–Merton knows of no such thing as the past or the future. When it models the stochastic process of the underlying as...
Automatic Differentiation for the Greeks The sensitivities of the value of an option to the model parameters, a.k.a. “the Greeks,” are crucial to understanding the risk of an option posit...
Product Risk Classification What is the PRC? The product risk classification (PRC) is a risk indicator that is based on quantitative models. It allows us to compare the fi...
The Volatility Smile Problem: From Within The Smile Problem is not a falsification of the BSM model ... The smile problem is produced from inside. In the following, I will say what the smi...
All Change SciComp Inc has been a major provider of derivatives pricing and risk models for two decades. Dean Tallam discusses the firm's outlook on technology i...
Monte Carlo Simulation of Long-Term Dependent Proc... This article briefly describes the Cholesky method for simulating Geometric Brownian Motion processes with long-term dependence, also referred as Frac...
Steering a Bank Around a Death Spiral: Multiple Tr... In this paper we start with the introduction of two pricing models to value contingent convertibles. One model ("rule of thumb") has its roots in cred...
Download Complete Article
Order Book Visualization
050317_bookmap_jan