Cowboy Escape

Cowboy Escape is a casual endless runner. The purpose of this project was to gain a better understanding on the design and implementation behind casual games, as well as practice data driven design techniques as part of a University module.

The general premise behind the game was to understand how data can be used to improve a game experience, as well as improve my scripting abilities along the way. I performed a general market analysis, and largely focussed on the mobile segment as these games are notoriously known for their live-service support.

In particular, I looked into games such as:

  • Temple run

  • Minion Rush

  • Subway Surfers

Due to the fact that all of them were within the endless runner genre, yet in a way they offered different experiences.

This prompted a very basic game design. Using available assets, the games theme "revolved" around a western cowboy setting - 3 track lanes and a series of pre-made, randomly generate obstacles.

In order to set-up and anticipate worth while quantatative data that would later be used to inform my iterative design, a series of game hooks were identified along side a speculative diagnosis that essentially anticipated potential problems that could occur.

With all the required tools in place, I created a general testing plan prior to any playtest sessions. The test plan was filled out subsequently to each session and pointed out the enquiry led by any changes.

The test plan can be found here: Link

In order to collect data, I researched different possible methods. While the most robust way would have been to utilise the built in system of Unity Analytics, for the purposes of my project this process would simply be too long (time wise).

Therefore, a proprietary script was used that collected the assigned KPI's and exported them in a sorted CSV. To further automate the process, I created a piggyback script that sends the files to an allocated e-mail address. While not the most elegant, I was able to centralise all my data and access it at any time.

With qualitative data in the form of questionnaires completed by testers after each test session, mixed with automatically collected quantitative data - I was able to bring them together into a centralised dashboard on google sheets which would be used to visualise and analyse my findings.

Based on these findings, I would perform certain changes to each test version of the game, collect data and compare the results to one another in order to observe the impact that those changes had. The process would be then repeated several times.

The final data report can be found here: Link

NOTE: The game is now under further development, in consideration of the data collected therefore there is a slight discrepancy between the most up-to-date changes and the evolution video. One of the changes based on the data collection, was that now obstacles are no longer spawned seperately to the tracks as this was harder to analyse with much more room for error due to randomisation.

Key Takeaways

  • Understanding of data driven design techniques, along with methods on how data can be collected, centralised and analysed.

  • Rapid prototyping, with data driven design changes.

  • Better understanding of casual and hypercasual games aswell as their respective markets.

  • Gained a better understanding on how to balance certain mechanics as well as implement them effectively. One of these being consistent random generation of a track.

Thanks for reading!