| ioi said: The whole reason why we don't quote a particular range or have something like your radio button ideas is because we can't. There is no way to calculate it other than to do a study and compare our data to other data over a certain period and use that as a base to produce a set of confidence levels and ranges. There is no mathematical way for us to calculate it, it would have to be done entirely empirically. |
Couldn't you do such a study continously? Would it be very time consuming to do it once and update it weekly? Or isn't there any reliable source that also publishes the results on a weekly basis?
| ioi said: It isn't a lack of wanting to communicate, believe me, rather that I don't see an alternative to how we do it at present other than to put everything togther from this thread and others onto one page and link to it at the bottom of every single chart we publish which seems like overkill to me when the majority of people (and certainly the majority of people in the industry) already understand. |
To gather and summarize the info from this thread and others onto one page seems a good idea to me. And I don't see the idea of link it to every chart to be overkill. A small and sober link at the bottom of every chart wouldn't hurt anyone or anything. But that's not the only possibility either. You could place the link at the header or footer of the site (but something that could be noticed as the place where the numbers are explained and that page would be exclusively to talk about that).
| ioi said: As I've said previously, I don't think we are acting any differently to any other similar service. These pages: don't make a point on-page about confidence levels, error margins etc |
Well, they must be doing something. Look at what Wikipedia says about them:
- "Media Create is a Japanese company that "gathers and analyzes data from the digital entertainment industry, specifically focusing on the Japanese console gaming market. Business operations include publishing, market research and consulting."[1] It is a popular website for people interested in learning the latest video game software and hardware sales figures from Japan [...] Media Create competes with Enterbrain's Famitsu and MediaWorks' Dengeki PlayStation in the market for providing Japanese game sales data. Because there are three different tracking firms, there will always be three different sales numbers for any software and hardware title. Which company to trust is a matter of debate, as none of the three major trackers are ever 100% accurate and whoever tracks the highest amount of sales for a given title fluctuates."
- "The Broadcasters' Audience Research Board (BARB) is the organisation that compiles audience measurement and television ratings in the United Kingdom. It was created in 1977 to replace a previous systems whereby ITV ratings were compiled by JICTAR (Joint Industry Committee for Television Audience Research), whilst the BBC did their own audience research. BARB is jointly owned by the BBC, the ITV companies, Channel 4, Channel 5, BSkyB, and the Institute of Practitioners in Advertising. Participating viewers have a box on top of their TV sets which tracks the programmes they watch [...] BARB's current research contracts have been awarded to three different market research companies: RSMB, Ipsos MORI, and Kantar Media. The contracts run from January 2010 to the end of 2015, with options to extend further. RSMB are responsible for survey design, quality control and calculation methodology. Ipsos MORI's role is to survey the characteristics of the television owning population, including an element of direct recruitment to the viewing panel. Kantar Media is responsible for establishing and maintaining the new BARB viewing panel. It has a separate contract covering meter panel installation, data retrieval, processing and audience reporting functions"
- "Box Office Mojo is a website that tracks box office revenue in a systematic, algorithmic way. Brandon Gray started the site in 1999. In 2002, Gray partnered with Sean Saulsbury and they grew the site to nearly two million readers when,[2] in July 2008, the company was purchased by Amazon.com through its subsidiary, the Internet Movie Database [...] The international section covers the weekly box office of 50 territories and includes historical box office information from three more, as well as provides information for box office results for individual movies from up to 107 territories. The site also creates an overall weekend chart, combining all box office returns from around the world, excluding the United States and Canada. The overall weekend chart currently tracks the Top 40 films as well as approximately fifty additional films with no ranking."
- "Famitsu is a line of Japanese video game magazines published by Enterbrain, Inc. and Tokuma. Currently, there are five Famitsū magazines: Shūkan Famitsū, Famitsū PS3 + PSP, Famitsū Xbox 360, Famitsū DS+Wii, and Famitsū Wave DVD. Shūkan Famitsū, the original Famitsū publication, is considered the most widely read and respected video game news magazine in Japan."
There isn't a single word about unreliable numbers or controversy. Now see what they say about VGChartz:
"While offering some information about their methodology through their website,[5] VGChartz does not publish any sources on how they get their data. Some sites, including Gamasutra and Wired News, have questioned the reliability of the information presented by the site.[4][6] VGChartz has defended the credibility and reliability of its sales data,[7] often comparing their numbers with the ones published by NPD Group,[8] although some charts have been retroactively adjusted to better match NPD's monthly reports."
Why do you think that happens?
| ioi said: I don't think they are intending to mislead readers that there will be margins of error associated, and neither is VGChartz. Do you know what error margins or confidence levels any of those figures represent? |
No, I don't. But then my question is: are their methods as controversial as yours? They seem to be tracking numbers, you do something different.
| ioi said: Barb for example (like Nielson in the USA) uses a panel of users with a box connected to their TV to estimate TV viewing figures in the UK. Their sample is around 11,000 people out of 60 million in the UK - less than 0.02% of the total population so their error margins would be on a comparable level to VGChartz. If we added a page similar to this - http://www.barb.co.uk/resources/reference-documents/guide-to-sampling-error?_s=4 - and linked to it at the bottom of every chart on VGChartz then would that help to make things clearer? |
That would definitely help, but they are just talking about Sampling Error. They also have this page - http://www.barb.co.uk/resources/reference-documents/how-we-do-what-we-do - and this FAQ - http://www.barb.co.uk/resources/reference-documents/faq
There needs to be a comprehensive explanation in my opinion and a FAQ helps too.
Prediction made in 14/01/2014 for 31/12/2020: PS4: 100M XOne: 70M WiiU: 25M
Prediction made in 01/04/2016 for 31/12/2020: PS4: 100M XOne: 50M WiiU: 18M
Prediction made in 15/04/2017 for 31/12/2020: PS4: 90M XOne: 40M WiiU: 15M Switch: 20M
Prediction made in 24/03/2018 for 31/12/2020: PS4: 110M XOne: 50M WiiU: 14M Switch: 65M
