Skip to content

msilb/auction-order-matching

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Build Status

Price-Time Priority Order Matching Engine for Auctions

This is an attempt to implement a simplified version of the order matching engines used in auctions (e.g. during opening or closing) on major exchanges.

Current limitations:

  • Only Limit Orders are supported

Auction Phases

Phase 1: Book Building

In this first phase an order book is open and accepts new orders. Orders may also be modified/amended or cancelled.

Phase 2: Price Determination Phase

Once Matching is initiated, order book is closed and no more orders are accepted.

First of all, orders are grouped and sorted by

  1. Side: Buy (Bid) or Sell (Ask)
  2. Limit Price: Buy orders with higher limit price and Sell orders with lower limit price have higher priority for execution (price priority). In case of multiple orders having the same limit price, orders which arrived earlier (time priority) have higher priority.

Next, total volume for each price point is determined on both sides of the order book (Bid and Ask).

Total volumes are summed up from the top of the order book to build aggregate supply and demand for each price level.

Finally, from the minimum values of supply and demand aggregates, the maximum executable volume is determined across all price levels.

Phase 3: Execution Phase

Provided the execution price and the execution volume determined in the previous phase, the orders can be executed. The rules specified above ensure that maximum of one order can be partially executed on either side of the order book. All other orders would either be fully executed or would remain open with full size.

Example

Consider the following order book before the start of Phase 2:

Order Bid Quantity Bid Price Ask Price Ask Quantity Order
B1 100 104.5 100.5 600 S1
B2 2500 104.5 100.5 400 S2
B3 1800 103 102 1500 S3
B4 500 102.5 103 1200 S4
B5 800 102.5 104.5 700 S5
B6 1500 99.5 -- -- --

Combining price levels from bid and ask, grouping orders and calculating total and aggregate supply & demand yields the following result:

Buy Orders Aggregate Demand Total Demand Price Total Supply Aggregate Supply Sell Orders
B1, B2 2600 2600 104.5 700 4400 S5
B3 4400 1800 103 1200 3700 S4
B4, B5 5700 1300 102.5 0 2500 --
-- 5700 0 102 1500 2500 S3
-- 5700 0 100.5 1000 1000 S1, S2
B6 7200 1500 99.5 0 0 --

Aggregation is done from top to bottom on the bid side and from bottom to top on the ask side (i.e. always from the best price on either side). To find the maximum executable volume, one needs to look at the aggregate supply and demand values across all price levels and find the max. For this purpose we need to consider the min of both aggregate values (supply and demand) since only so much can be satisfied for any given price level. In this case our auction price is determined as 103 and our max execution volume is 3700 (=min of 3700 and 4400).

Once these two values have been determined, we can go back to Table 1 and proceed with order execution:

Execution Price: 103

Order Execution Bid Quantity Bid Price Ask Price Ask Quantity Execution Order
B1 100 100 104.5 100.5 600 600 S1
B2 2500 2500 104.5 100.5 400 400 S2
B3 1100 1800 103 102 1500 1500 S3
B4 0 500 102.5 103 1200 1200 S4
B5 0 800 102.5 104.5 700 0 S5
B6 0 1500 99.5 -- -- -- --

About

Simplified price-time priority auction-style order matching engine

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages