• Akira Toriyama passed away

    Let's all commemorate together his legendary work and his impact here

A case for using more statistics; eventually.

ggx2ac

Member
Expert
(They publish Dark Souls)
I want more statistics; I don't mean more line/bar charts. For example, from my webpage in the Nintendo section there would be charts at the bottom of the page. I have said before that I am not a big fan of it because all that data is already in the tables above so it's redundant, I want to find other statistics to apply there.

I haven't done any (significant) statistics for several years so it will take some time for me to demonstrate the value of it. It is very relevant in a place such as here where data is highly valued. I understand that Statistics is not just something you can walk into like making charts. I do not have any resources yet that would help teach you it, maybe there will be AI you can use that you can tell it to do statistical tests and then it becomes as easy as making charts.

What is (the Wikipedia definition of) Statistics?
Statistics is a mathematical body of science that pertains to the collection, analysis, interpretation or explanation, and presentation of data.

With statistics, you observe something that would make you want to come up with a hypothesis. You then have to apply some statistical testing to disprove your hypothesis. That testing process requires a data sample taken from a population because you are unlikely to have data from the whole population but that's okay, you will calculate things such as the standard error to account for those issues with your sampling.

There's too much to go on trying to explain all this. The point is, I have ideas I want to test and unfortunately, they're not so simple to test thanks to how difficult it is to get data from the video game industry.

Here is a past example, I made a thread on an old forum (I am not going to repost here everything I said from there) with the thread title:

Double Dipping is not a thing that Publishers aim for when they make late ports, it's just your gaming/buying habits


This is where pushing an idea with a lack of data can really backfire on you however, I did not come up with the idea of double dipping and publishers certainly did not either. Those on forums that use the term themselves did not have data backing up that publishers would make late ports just to get you to buy a game again on another platform.

While I do not have data still to show that publishers don't make late ports on other platforms to make you buy the same game again, it's very easy to think about the probability needed to get a significant number of sales solely on double dippers by making late ports. You need: The probability that a person owns the console/PC required to play the game you release, the probability that they will purchase that game, the probability that one or two years down the line when you release a late port that the same person will own the other console or PC required for the late port you release and the probability that they will end up purchasing that late port. It's a lot of steps that it is very unlikely that publishers are sitting at a table thinking, "We're going to release the same game again on another platform in the hopes that everyone double dips."

Moving on, it would be great to come up with some hypotheses to test if it wasn't so difficult to acquire data from the video game industry.

There are two hypotheses I can think of right now that I would want to test, one I can remember coming up with years ago is the idea that a same-console generation video game sequel will sell less than its predecessor that released in the same console generation. I would need a population of games and its sequels that released in the same console generation to make those comparisons, there are game series that probably aren't affected by this so the test would be to find if a significant portion of sequels end up selling less than their predecessor that released in the same console generation.

The other hypothesis I want to test is, "The top 5 video game companies have a larger share of the market than ten years ago." This would require knowing how much sales total have been made worldwide in 2022 for example and then comparing that to ten years ago and it is not so simple but, it is definitely something that would be useful to know. If the top 5 video game companies have a larger share of the market than ten years ago then that means there's not going to be much growth available for other companies in that space, so to compete in that space you would then need a big company to enter, think when Sony entered the console market when it consisted of Nintendo, Sega, SNK, Hudson. Now for big companies you would need something like Google (and we saw how that went, remember Stadia?) or Amazon to enter that space.

This is why I am making the case for more statistics; I understand it is something that most will not be able to participate in compared to making charts. I am showing you that it is possible to do something more with the data you have, to test something you have observed and then show it off to others. The only issue is how secretive the video game industry is with that data that it makes to more difficult to test something.
 
Stats are indeed great but like you mentionned, you need a large and extensive database.

Since we are unlikely to subscribe to a tracker, that means enriching the database yourself. For that, I believe the JP market is the easiest to work on.

Manual work will always be needed, especially when you want to make different comparisons. More broadly speaking, the dataset is the key element into any statistical work.

For instance, testing "Does one console have an impact on the other ?" would need you to run a correlation test between hardware sales. But then, you'll notice seasonality and try to remove it (either by spliting your dataset into similar months/periods) etc.

Lots of possibilities, but the sine qua none remains to build a comprehensive database.
 
Back
Top Bottom