Developers

To further reduce latency on shipping data to the client, we have reduced each streams message payload to the bare minimum and allow the client to decide if they want messages decorated with optional data.  

For example, for odds streams we do not ship the Country, Competition or Bookmaker names but do provide the IDs as this information is available via the reference data streams/endpoints. However, the client can ask for these names to be shipped if preferred.  

As new derivative markets are continuously added by bookmakers we have built a new method to classify odds markets into a unified representation called SuperOddsTypes to avoid issues encountered when mapping traditional odds types to allow clients to subscribe to only the specific markets (or subset of the market) that they require.  

For example, a Soccer Over Under market can have a period i.e. full time, 1st half, 2nd half, extra time plus various selections for both teams, individual participants plus the line value ladder or ranges and so on. SuperOddsTypes can be queried within a subscription with AND, NOT, OR to return only what is required.  

Fusion utilizes Server-Sent Events (SSE)/ Eventsource as a transport mechanism to stream the odds. This uses regular HTTP requests for the persistent connection, not a special protocol greatly simplifying the integration process.  

Connections are initiated by the client, and once connected the server will send events in text/event-stream format without closing the connection. Similar to WebSockets in that they happen in real time, SSE are a one-way read-only communication method from the server.  

Each payload message in the stream contains the event type which can be `live`, `context` or `heartbeat` to allow the client to process messages via different callbacks/listeners and act appropriately i.e. heartbeats can be used tp monitor the status of the subscription.  

Each event also is decorated with SSE LastEventId which allows clients to automatically re-connect if the connection is interrupted and the data stream can be continued from the point of interruption, so no messages are lost. Clients also have full control so they can decide not to use the LastEventId feature and use timestamps on reconnect to gap fill if they so wish.  

It is possible to have many streams open at the same time with the SSE streams being multiplexed over the same physical connection. Highly available, resilient and load balanced Odds Distributor endpoints are available from which to consume data and are accessing via a single URL and the subscription is transparently placed onto the least loaded endpoint.  

There are many SSE/Eventsource libraries available for all major languages and at least getting started to consume Fusion data for review is a lot simpler than using legacy messaging protocols.   

We do also provide documented example code in Java, C#, Javascript and Python to aide integration on the client side.  

TXODDS has several different legacy feeds such Trader Feed/AHC, Market Feed/xml2 & Push that can deliver information about competitions, fixtures, and odds.   
Fusion uses a new platform to manage this reference data but to protect existing customers investment in the Trader, Market and Push feeds the IDs generated by these systems are optionally available in the Fusion feed via the IOI feature described above.   

Users of Fusion will therefore be able to obtain all required information to support the odds subscriptions from Fusion itself without having to rely on other feeds. The customer experience and integration effort will be greatly simplified as all feeds will be delivered in a single, consistent manner.  

The service allows a client to have one or more subscriptions concurrently, with each subscription’s scope (or interest) expressed in a very simple subscription DSL. The client application communicates with the server-side Odds Distributor component that authorises and manages all the current customer subscriptions.  

For example, clients can query to get just pre-game or just in-play pricing or both via the same endpoint. TXODDS have the same FixtureId for an event whether it’s pre-game or in-play but some customer systems can require pre-game to be forwarded to one internal endpoint and in-play to another. Clients can have multiple subscriptions to achieve this without having to read the payload data.  

The same applies to sports, bookmakers, competitions, SuperOddsTypes so if a client wants to have different subscriptions with different bookmakers and SuperOddsTypes for each sport or competition that’s also possible.   

With the entensive global coverage TXODDS provide for specialist bookmakers, clients can choose the best in breed for their pricing, whether that be Asian bookmakers for Soccer, US bookmakers for NBA, NFL, MLB,NCAA and European bookmakers for Eurobasket it’s all possible.   

There’s also no need to `book` an event or fixture in Fusion, all events are available based on the scope of a subscription.  

The Fusion endpoint(s)  are available only via HTTPS to provide encryption to ensure your requests are being sent to our servers and to protect the privacy and integrity of the exchanged data from malicious attacks or being read by a third party.   

Customers will need to supply external facing IP address(es) from which they will connect as these will have to be added to the allow list on the firewall. Customer subscriptions will also be required to supply credentials for authentication and access to the feeds.  

Compression can also be requested by the client to further reduce bandwidth consumption and decrease latency.  

You can request only updates to the odds you are interested in, so no need to refresh the entire catalogue of event(s) constantly thereby reducing your server bandwidth and processing resources.   

As mentioned earlier using SSE libraries and the LastEventId when the client is running it can automatically and transparently handle short interruptions like network failures and distributor restarts/maintenance and request any updates including market OTBs during the time of the disconnection.  

For longer outages/clients being restarted on the customer side if the LastEventId and/or the last received Ts ( timestamp ) will need to be maintained/stored on the customers side so you can request `as much or as little data` as required.   

We send our server timestamp which is set always to GMT to you to ensure that there are no time zone or clock synchronization problems. Wherever you are in the world or whatever time your local PC or server is set to will not affect the data we send you. 

• A blend of bookmaker’s prices for core markets based upon a variety of factors leveraged by analysis from our market intelligence and historical data product set. 
• Stability/availability – from our analytics we track and report on the bookmaker’s uptime i.e. frequency and length of pricing and OTBs plus the source website/feed availability metrics. 
• Coverage – which bookmakers also cover the most events and markets.
• Market experience – the global betting know which bookmakers’ price `accurately` and are less prone to errors or delays in updating / repricing following goals/ in play incidents. 
• Best of Breed – use different blends of bookmakers per sport.
• By knowing where the market is pricing a match customers can track whether their own pricing is `in or out` of the market in real-time by re-margining based on their models and act accordingly 

Features 

• The stable price will be available no earlier than 3 days prior to the fixture although updates are processed up to 8 days before the fixture. 
• Baseline price availability – i.e., we set a limit to say we must have X out of Y bookmakers to produce a price for further calculations to be made and that these prices are within the defined time windows for pre-game and in-play. The time windows vary (shorten) depending on the proximity of the fixture in question. 
• The Baseline / Incoming prices are then demargined.
• They are then converted to implied probabilities.
• We check there is a consensus i.e., supremacy probabilities are within a defined range and are similarly aligned otherwise we send an OTB. 
• Further defensive measures are applied to remove outliers and arbitrage opportunities and ensure a consensus. 
• Consensus probability is then converted back to a price and emitted as the StablePrice.

Markets currently Supported 

• Soccer core markets – Asian Handicap, Over Under/Totals, 1X2 for both full-time and first-half market-periods 

Fusion synthetics allows customers to define a consensus view of the market based on subscriptions to underlying bookmakers. Certain bookmaker prices may be considered better than others for a particular application and synthetics allows you to define a formula to create a new synthetic bookmaker price based on these chosen bookmakers. You can subscribe to this synthetic bookmaker and when any of the chosen bookmaker prices change, the synthetic bookmaker’s consensus price will be recalculated, and updates delivered in the same way as for a standard bookmaker. It is possible to add a weighting to bookmakers to reflect their relative importance in certain modes of operation. Conditions for delay periods and probabilities can also be added to the definition to further control the offers that are admitted into the consensus allowing a fine degree of control over the pre-game and in-play synthetic prices that are generated.