#	Pagina
attuale pagina	/open-h2020/projects/210428/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - Gamebooster (Promoting the Efficiency of the European Online Games Industry through an adaptive gaming platform and portal.)

Teaser

Summary

The European games industry is characterised by its reliance on casual and free-to-play online games. Since this is a growing but chronically underperforming sector, the gaming industry in Europe is not meeting its full potential. Games are the third largest entertainment medium worldwide by market share, behind only television and gambling. Poor leveraging of the medium is common throughout the EU. The European Social and Web-based Games industry is weakening in the face of venture-capital and corporate-backed opposition from the dominant markets, the USA and China. These markets are expected to dominate the casual/social/mobile games market, making up 50% overall share by 2020. This will effectively limit the underperforming European development industry to retreat to specific niches within the market. Current social platforms, ecommerce platforms, and other major online services indiscriminately collect data (e.g. Google, Facebook, Twitter, Amazon) and store it in a mostly unsafe manner (e.g. the Yahoo, DropBox, TJ/TK Maxx, Cardsystems Solutions Inc., AoL, Department of Personnel Management, and LinkedIn breaches). Worse still, much of this data is sold in a â€œfirehoseâ€ format, where everything is sold to the user with no regard for its use. This supports illegal data gathering and profiling by state entities, and unwanted solicitation from social or commerce platforms like Twitter and Amazon. Due to this haphazard data collection practice, two contrasting problems emerge. Highly-targeted advertising causes users to become justifiably distrustful of platforms which collect everything about their lives. And targeting is also difficult and expensive, and reliant on Big Data and Deep Learning systems to trawl through the data to find any meaning. This in turn leads to poor recognition of trends and poor optimisation of systems, due to the inaccurate mapping of diffusion through player networks, such as the sharing of Facebook games.

The Gamebooster platform carries out automatic and user-driven optimisation of the game experience, from mechanical gameplay parameters, to the platform itself prioritising games that garner positive feedback from our user base. This is achieved via machine learning algorithms that identify user behaviour that indicates a quality game, and typical user preferences for gameplay. Our development plan intends to apply Big Data and Deep Learning systems in a more targeted manner than that used by other platforms. Our platform focuses on HTML5 games rather than the older, less-secure, and proprietary Flash package. By employing this protocol, games published to this standard are inherently capable of being played on multiple platforms and even being embedded in instant messaging services such as Facebook Messenger. Where we truly add value is combining analytics with the following traits â€“ limited connection through social networks, cross-platform compatibility, and repeat visits. This allows the tracking of network diffusion of games and products, without the â€œnoiseâ€ in a system that is simply attempting to capture as much data in as possible in its dragnet. We break away from the hazardous and unethical revenue models of other social game delivery platforms like Facebook, the Google Play Store, and the AppStore. They rely on indiscriminate harvesting and sale of user data, and the unfettered serving of ads which exposes their end users (and true source of revenue, not the ads they serve) to significant risks. This includes the loss of personal data through breaches, or hostile/unethical actors simply buying access to the â€œdata fountainâ€, and malware through advertisements. As such we not only improve the gaming experience, we reduce the danger our users face when innocently enjoying their free time.

The general objective of our project is to expand our platform to full-scale functionality with respect to the optimisation and recommendation engine, the Big Data and Deep Learning platform

Work performed

In early-mid 2017, we carried out a Feasibility Study under the auspices of the Horizon 2020 SME Instrument Phase 1. In this project we assessed and addressed our in-house capacity to carry out the activities we need to fully commercialise our platform with the unique, innovative controls and systems. We established our goals and criteria for success in accordance with the SMART (specific, measurable, attainable, realistic, and time-constrained) system and addressed the hybridisation of DMAIC (define, measure, analyse, improve, and control) techniques which our system itself will be applying to data, and our own agile product development process. Furthermore we established the business interest from dealing with advertisers and other customers, resulting in a projected 300% increase in our revenue-generated-per advertisement.

Final results

Dynamic optimisation of the platform and certain gaming parameters is intended to improve the experience of both the game developers and the gamers. By allowing the user to adapt a gameâ€™s settings on the fly, or letting the platform guide these improvements, we bring the mobile and social gaming experience closer in line with the AAA gaming industry where the user is given greater scope to adjust difficulty and other aspects of the gaming experience. For developers our platform optimises search functions and recommendations, rewarding a greater level of diligence in the production of social games. This results in a positive feedback loop where games with long session durations, high replay rates, and other traits currently being established as signs of a high-quality user experience, are given priority in search results and recommendations. Further, for gamers, we protect their data and respect their privacy.