"Kartlytics" to analyze Mario Kart's play data and explore the law of victory

ByJohn Flinchbaugh

Nintendo'sMario Kart SeriesSpeaking of it, it is a popular racing game boasting over 90 million sales worldwide. Mario Kart 64 which is the second in this seriesNINTENDO 64As the game of the domestic shipments are the most number of shipments, there are still a few fans to play. There is a project that attempts to unlock the winning method of such Mario Kart 64 using big data and analytical tools.

Kartlytics: Summary
http://kartlytics.com/

Dave Pacheco's Blog »Kartlytics: Applying Big Data Analytics to Mario Kart 64
http://dtrace.org/blogs/dap/2013/08/08/kartlytics/

Dave Pacheco who did this project originally said he was playing Mario Kart 64 with his colleagues, "Is it appropriate for the handicap to give non-powerful items to players who are in good rank in the race?" , "How much does it affect the final ranking for the first lap", "Which character is the character that can win the most", and so on, interest in Mario Kart 64 seems to be raised.


And again and again many times "Which character of lightweight or heavyweight is better" or "How much time do the bananas and these get hit or falling out of the course" and others with friends Although he seems to have repeated discussions, he finally got video capture of how he was playing Mario Kart 64, and he came to finally analyze the contents of the race.


The race data analyzed from the project start to the present can be confirmed on the following page.

Kartlytics: All races
http://kartlytics.com/#races


Links to all 239 races are affixed from this page, so if you click a date, for example ...


It jumps to the page where detailed race data and race movies embedded in each date are embedded.


Here is a record of each player's score for each hour ......


Events that occurred during the race are written with capture photos.


The analysis result of this accumulated data is summarized in the summary page and top page of kartlytics.


For example, an item box that gives players items to raise lace. For each item that you can get from now, the probability is shown on which ranking items you can get with what probability. There may be experiences that triple mushrooms appeared when running in the lowest place and thought that "I want a little better item ...", but the probability is considerably higher when I look at the analysis result.


This is the probability of the item that the player of each rank can get when touching the item box with one item possessed.


Furthermore, the case of Keithings, which falls from 1st place to 4th place within 5 seconds, happens in more than half of races in 239 races. Even if I ran first place, I will be the lowest in the moment ... ... I can say that Mario Kart 64 seems like a result that seems not to know what happens.


The reason why we are doing this analysis isJoyentEmployees. Joyent is a company that provides cloud facilities like those used in real-time web and mobile applications, and it is a cloud storage service that simplifies processing of big dataMantaHas recently been released, it seems that it was a project that also implied the confirmation of the operation of this manta.