Granularity order book model

We continue our series of posts on order book models by considering an order book model proposed by Cristelli et al. [1], which is able to reproduce power-law price change. Originally this model was develop to study order impacts on a price in case of finite liquidity (in case when the order book is not tightly packed). Indeed it seems that when there are noticeable gaps behind the best bids, going deeper into the respective order book side, a couple of consecutive orders may cause power-law fluctuations.

The model

We initialize this model by filling it with order uniformly distributed around best bids. The original paper did not specify any specific model initialization rules, hence we picked on our own. So please allow certain burn-in period for the model to allow for its true dynamics to emerge (around 5000-10000 time ticks should be enough).

The price in this model is defined as an average between the best bids:

\begin{equation} p(t) = \frac{a_0(t) + b_0(t)}{2} , \end{equation}

here \( a_0(t) \) and \( b_0(t) \) are best ask and bid quotes (prices associated with the order) at certain time \( t \) respectively. We took liberty to define the price, in case of any side of the order book would be emptied, as the last executed order price.

In this model quotes are defined in integer price ticks relative to some reference level (price at "zero" level). No quote may be expressed in fractional price ticks. In this sense price tick is smallest unit of price. In the original paper it is set to 1 (without loss of generality).

The time is also discrete and runs in time ticks. An order is created at each time tick after the model initialization. The order is either ask or bid with equal probabilities. The order is market order with probability \( \pi \) or limit order with probability \( 1 - \pi \). In the original paper \( \pi \) is estimated to be roughly \( 0.32 \).

If the order is limit order its quote is assumed to be uniformly distributed in

\begin{equation} [ b_0(t)+1 , b_0(t) + k s(t) ] , \end{equation}

if the order is ask order, and uniformly distributed in

\begin{equation} [ a_0(t) - k s(t) , a(t)-1 ] , \end{equation}

if the order is bid order. Namely the ask orders are put at least one price tick away from the best bid quote and vice versa. The order are put at most \( k \) times spread (defined as difference between best ask and bid) away from the opposite best bid. In the original paper \( k \) is estimated to be between \( 3 \) and \( 4 \). If order book would be empty on the respective side, if \( b_0(t) \) or \( a_0(t) \) would be undefined, then the order would be put one tick away from the last executed order price. This is once again our interpretation, which is not present in the original article.

If the order is market order, then it simply triggers the best opposite quote.

This model also includes simple order cancellation mechanic. Each order is canceled after \( \tau \) time ticks. The original paper argues that for this case most realistic results are produced when \( \tau \in (200,750) \).

In the applets below we also use parameter \( \delta \) which describes both number of time ticks between the visual updates and used to define return (price difference):

\begin{equation} r_\delta(t) = p(t) - p(t - \delta) = \Delta_\delta p(t). \end{equation}

Interactive applets

To understand the model and its dynamics better you should study the interactive applets below.

The first applet, as has become usual, shows us how the structure of the order book evolves as the time goes on. Here we have assumed that the reference level is at \( 100 \) monies and the price tick is set to \( 0.1 \) money. This assumption were done simply to obtain easily understandable figure.

The second applet is in certain sense traditional Physics of Risk applet for all financial market models. Though there are some specific differences related to the model in question.

First of all instead traditional price, or log-price, series plot on the bottom left you see relative price series plot. Recall that this model operates using distances in price ticks from some undefined reference level. To get price series you should add some fixed number (price at "zero" level).

Another difference is related to how return is measured in [1]. In the paper return is measured as absolute difference between the relative price at two different points in time. To differentiate between the return in traditional sense (difference of log-price at two different points in time) we call this return simply "price change" and use \( | \Delta p | \) notation.

Note that \( | \Delta p | \) is power-law distributed when \( \tau \) is small. Also note that in order to match the original paper results we have defined the price tick as price measurement unit (in other words price tick is equal to \( 1 \)).

Acknowledgment. This post was written while reviewing literature relevant to the planned activities in postdoctoral fellowship ''Physical modeling of order-book and opinion dynamics'' (09.3.3-LMT-K-712-02-0026) project. The fellowship is funded by the European Social Fund under the No 09.3.3-LMT-K-712 ''Development of Competences of Scientists, other Researchers and Students through Practical Research Activities'' measure.

References

• M. Cristelli, V. Alfi, L. Pietronero, A. Zaccaria. Liquidity crisis, granularity of the order book and price fluctuations. European Physics Journal B 73: 41-49 (2010). doi: 10.1140/epjb/e2009-00353-6.