4 suggestions on publishing sales data:

1. preview numbers published (perhaps in an article rather than in the charts) ASAP within a day or so of the end of the sales week: Hardware - the expected range (e.g. console X 120-150K range, console y 105-130 range, console z 30-50K range). Top 10 software, but no actual sales numbers.

2. preliminary numbers published in the charts not official numbers do not cause LTD totals to change in the home page bar graph graph or in the tools line graph, published on the same day every week (Thurs/Fri/Sat? or earlier if ready). Most probable number (MPN) for hardware. Top 30(?) only MPN software with numbers.

3. Interim final numbers replaces preliminary numbers, are the official numbers for the week until future adjustments. Subject to potentially several rounds of adjustment. Full software chart. No specific timeframe for publishing this data, there could be multiple weeks of charts that are sitting at step 2 if there is a lot of verification and adjustment work that needs to happen before interim final numbers are published.

4. final numbers the numbers for the week that will not be subject to future change through adjustments, would normally be at least 1 year before there can be reasonable confidence to make numbers final.

Step 2 can be skipped if there is sufficient confidence to go straight to sept 3 on the due day for step 2.

Bans for anyone who criticises numbers at step 1 or 2. They are rough figures and people should see them for what they are, just an indicator of where early data is indicating sales should land.

Advantages: predictable dates for publishing unofficial data. Transparency of how much stock people can put in the reliability of numers (i.e. only step 3 and 4 data can really be used for analysis), but it still gives us datahounds something to talk about. Can take bets on whether step 3 numbers will flip positions if step 2 data shows very close numbers.

Disadvantages: for us none that I can see. For the VGC team, a bit more work, but hey what do we pay them for?... err I mean what do the advertisers pay them for?

And for the moderators that have to deal with the epic fallout - no one pays us anything

Preview numbers were already tested and failed here in the past. As an example, the old Japan preview was the worst. Not even in the same ball park for estimations when compared to Famitsu etc.