Nav: (Display/Hide) - Home - About the Author / this page

Current Projects: Americana Engine (Game Engine Development)

Saturday, January 7, 2012

Evolution of Leaderboards: Product Development

Many weeks ago (as early as last year), several players decided to step it up and create their own leaderboards based off their high score tables, and eventually groups had players designed to gather scores of players who wished to participate in the group leaderboard, since the board was for friends only. Fast forward to December - my definition of 'leaderboard' was taken a step further when the friend requirement was dismissed with the introduction of a true 'top 200' leaderboard as proof of concept as first, then a regular recurring series after many positive reactions.

I do it mainly to satisfy the needs of the players who want to know how they rank up. But honestly, it seems more like product development.

Prior to the creation of the board, I was looking through posts on forums and this group and discovered that several people were interested in a group-wide or national leaderboard. I went out to develop something that would solve this problem, as well as bypass the 'top 50' on a local leaderboard.

Above, we have the comparison of the first leaderboard that was hosted on this blog with the discovery of the score parser (basically, this is what it actually does - take raw bits of information and parses them into readable data players can interpret easily). More tweaks were made and released on Facebook to my ZB friends for feedback.

A week later the leaderboard went live on the official add me group, with mostly positive feedback, but several had concerns about how this data was retrieved, which was answered. I later went on to add a few more features that I determined might be helpful such as person's level (in order to show true skill as mastery levels weren't factored into score) and general score statistics. One additional feature, ranking arrows, were planned for release next week, but I'll let the viewers decide that change.

A side note on stats: there is a field not listed here called "Percent of Scores Over" - I like stats like this, but it can also be used to determine approximate board difficulty (how easy it is to get points) - a peak on lower level scores (like over 500K or 1M) may indicate the board is easy, peaks on scores lower than this might indicate the board is harder, and low minimum score percents can indicate that difficult multiple-gap techniques might be in play on this board. This can be used as a baseline to determine if a unique (possibly user-created) board has the potential be used again.

The snowpocalypse board had for minimum scores 74-54-45-25 for position 10, 30, 50, and 100th place for the group respectively in terms of percentage of the top score on the board. This indicates that speed and luck (time balls) was more prevalent than skill. Major Mouthful, on the other hand, has 53-38-28-14. This represents that a high scoring technique was more likely used than just speed (in this case a double gap strat was used here).

What can I summarize about making this board? If this was a product in development, it's going through its incremental cycle right now.


Tim said...

I have a couple of questions.
First, how do i get on this leaderboard or at least connect to the leaderboard to see all of these high scores.
Second, can this leaderboard be segmented by level. Example would be the highest scores at each level. Thanks!

Brendan Chan said...

1. Since I personally manage this leaderboard, which is a group-wide leaderboard, add yourself in this group and I will place you on the leaderboard shown above starting next Tuesday.

In the future I will probably create a page where you connect using facebook and will grab only the necessary information (name, facebook id) to place you in one. However I currently don't have the funds to obtain dedicated machines that will handle leaderboards with potentially tens of thousands of people at this time.

2. It is possible to be segmented by level, however scores are NOT realtime (I can only get results from previous boards) so there would be no way to tell exactly what level that person achieved a high score on.