Finishing my Scientific Methodology course has made me curious about data, simulations and R. In addition, a mix between working with the book "Hands-On Programming with R" (by Garrett Grolemund; ISBN 978-1-449-35901-0) and playing video games since the sweet age of five(ish) has brought up the idea for this blog (e.g. this post). My main objective is to learn R and data handling in order to (maybe) become a data analyzing super psychologist (currently studying in Sweden; this is an official title). On my first post, I will take the racing game Gran Turismo Sport serious, deadly serious.
Gran Turismo Sport
Up to now, everyone around the world of the has heard about the controversial content of Gran Turismo Sport, where you can race various cars on various tracks with various laps and various opponents, day in and day out. The legend has it, that you have to be online all the time and that something called "career" was banned from the system, leaving the die-hard fans of the previous parts (well, only 1-5; I still remember spending a lot of nights on part 2, hunting for milliseconds) with a sore feel in their tummies. In a salty addition, cars and tracks were trimmed so you can't expect to drive your real car in this game (unless you have made an early Bitcoin investment). Still, the free open beta had a total player base of one million, and launch sales were pretty good. In total, that means that a lot of players are competing each day, driving a lot of kilometers (or miles) in their pretty cars, using a lot of fuel to keep them going.
The annoying thing with co2
Yep, you read right, co2 emissions are actually a thing and they are not made up by the Chinese. It is a serious issue of this world and this society and it's not getting better big style. Looking at Germany, out of the 906 million tons of produced co2, 160 million tons came from traffic. 160 million tons equals roughly 2.6 billion alpacas or (also roughly) 100 million hippopotamuses (yes). So, co2 is produced from traffic, Gran Turismo Sport is a racing game where you drive cars, cars are traffic, Gran Turismo Sport is traffic, Gran Turismo Sport equals co2, easy or in other words: stop playing Gran Turismo Sport.
By the way, if you want more (serious) information about the topic, Sam Harris has a good podcast about it (link).
Some methods and data collection
To give this thought a bit more ground, I collected the track length of every track presented in the game (in km). The track data was collected via this video. The next step was the collection of the co2 emission per km of the car collection. There was the first wall I hit, some of the cars are prototypes, some of them are high-performance race cars with modified engines and rare data on the emissions. Therefore, I decided to only collect data for the cars of Group N, the "lay people cars". This data was either available through Wikipedia or the manufacturer's page (and sometimes hidden, thank you Toyota and Asian cars in general.). In total, there are 26 tracks with the shortest track being Northern Isle Speedway (0.9 km) and the longest being the Nürburgring Nordschleife (20.832 km), and 33 cars with the Peugeot 208 GTi having the lowest amount of emission (125 g) and the Aston Martin One-77 (hell of a "lay people car") having the highest amount (572 g) of co2 per km. With those data sets, I was ready to think about a way simulating co2 emission per player per time period.Qualifying in R
The book I mentioned in the beginning, has lead me to make a weighted dice, a card shuffling and dealing simulation, and right now a slot machine program. It is written in a good and light understandable way, motivating to turn the page and dive into the handy world of R. Coming from those training rounds, I gathered some functions and replicates together to form my simulation routine. First, I loaded the Excel library and both sheets into RStudio, using the following bit of code.
Next step was creating a copy variable of both sets, saving cars in CARS and tracks in TRACKS (creative). Now we are heading to one of my favorite functions in R, the sample function. Since it is not likely that every player will drive only one car on only one track, randomness comes onto the stage.
Next step was creating a copy variable of both sets, saving cars in CARS and tracks in TRACKS (creative). Now we are heading to one of my favorite functions in R, the sample function. Since it is not likely that every player will drive only one car on only one track, randomness comes onto the stage.
With the extraction of the numeric values "co2km" and "lengthinkm" (happening with cars$co2km and tracks$lengthinkm), a calculation of the co2 per lap can now be arranged. We now have 1000 random drawn cars (with replace) driving happily on 1000 random drawn tracks (also with replace). The following snippet calculates the mean co2 emission in kg.
Since it is hard to estimate the number of laps and opponents being chosen by the player, I took advantage of the fabulous feature of daily races, presented by the Polyphony Digital team. Daily races are three event like races on three tracks with three different car groups. Usually, those races are online races, against human players. For reasons of simplicity (since a calculated, simulated human could race another and it would no longer be a single easy case), I only simulate races against the AI. Taking the daily races (data coming from reddit user smthng) from the 19th of October up until the 15th of November into account (57), I gave both, the laps (ranging from 2 to 20) and the number of opponents (ranging from 12 to 24) different probabilities based on their appearance. In R, this looks somewhat like this:
So yeah, we now have a rounded mean value of around 7 for the laps and 16 for the opponents. Using this code, one player (alone on the track) would produce an approximate amount of about 9.5 kg of co2 per race. But hey, since we are not alone in this world, we have to take the other cars into account too.
Great, this gives us a mean of 156 kg/co2 per race. Taking this value times 57, we have the amount of co2 produced by the daily races between the 19.10 and the 15.11: 8892 kg or two male Asian elephants or 14040 kg a month (taking it times 90; 3 races a day, 30 days a month). That is 1.5 times more than the produced per head amount of co2 in Japan within a year (2015). And that is just one player racing the AI. If we only take the sales from the release week (150286 units in Japan) and assume that these players are daily racing enthusiasts, they produced the nice amount of about 2 million tons a month. And still, we are only talking about daily races.
I like to think of release purchasers as one of the most motivated groups to put time into the game. Let's say that they are participating in 25 more races a day in addition to the three daily races, making it a total of 28 races. With this racey attitude, those Gran Turismo Sport loyalists are dooming Japan's environment with a total amount of 19.8 million tons of co2 per month, adding the insane amount of 237.6 million tons to the yearly co2 emission (which was 1.2 billion back in 2014). That is the amount produced by traffic (land and air) by the whole population of Japan back in 2008. With the goal of reducing the emissions on a yearly basis, this is a cough surprise, right?
Since it is hard to estimate the number of laps and opponents being chosen by the player, I took advantage of the fabulous feature of daily races, presented by the Polyphony Digital team. Daily races are three event like races on three tracks with three different car groups. Usually, those races are online races, against human players. For reasons of simplicity (since a calculated, simulated human could race another and it would no longer be a single easy case), I only simulate races against the AI. Taking the daily races (data coming from reddit user smthng) from the 19th of October up until the 15th of November into account (57), I gave both, the laps (ranging from 2 to 20) and the number of opponents (ranging from 12 to 24) different probabilities based on their appearance. In R, this looks somewhat like this:
So yeah, we now have a rounded mean value of around 7 for the laps and 16 for the opponents. Using this code, one player (alone on the track) would produce an approximate amount of about 9.5 kg of co2 per race. But hey, since we are not alone in this world, we have to take the other cars into account too.
Great, this gives us a mean of 156 kg/co2 per race. Taking this value times 57, we have the amount of co2 produced by the daily races between the 19.10 and the 15.11: 8892 kg or two male Asian elephants or 14040 kg a month (taking it times 90; 3 races a day, 30 days a month). That is 1.5 times more than the produced per head amount of co2 in Japan within a year (2015). And that is just one player racing the AI. If we only take the sales from the release week (150286 units in Japan) and assume that these players are daily racing enthusiasts, they produced the nice amount of about 2 million tons a month. And still, we are only talking about daily races.
I like to think of release purchasers as one of the most motivated groups to put time into the game. Let's say that they are participating in 25 more races a day in addition to the three daily races, making it a total of 28 races. With this racey attitude, those Gran Turismo Sport loyalists are dooming Japan's environment with a total amount of 19.8 million tons of co2 per month, adding the insane amount of 237.6 million tons to the yearly co2 emission (which was 1.2 billion back in 2014). That is the amount produced by traffic (land and air) by the whole population of Japan back in 2008. With the goal of reducing the emissions on a yearly basis, this is a cough surprise, right?
Finishing and concluding
Those estimations and the (real) data of Japan's traffic emission are leading me to the ultimate conclusion, that you should stop playing Gran Turismo Sport (or other racing games, please) immediately.
No comments:
Post a Comment