This is from the first Japan predictions.
average | actual | average accuracy | |
 | 104,576 | 97,723 | 92.99% |
 | 93,924 | 69,511 | 64.88% |
 | 41,080 | 39,676 | 96.46% |
 | 52,980 | 69,655 | 76.06% |
 | 7,617 | 6,108 | 75.29% |
| my prediction | actual | my accuracy | ||
| 115,000 | vg$ 5 | 97,723 | 82.32% |
| 95,000 | vg$ 5 | 69,511 | 63.33% |
| 45,000 | vg$ 5 | 39,676 | 86.58% |
| 65,000 | vg$ 5 | 69,655 | 93.32% |
| 6,500 | vg$ 5 | 6,108 | 93.58% |
All you need to know is if you had a better accuracy than the average then you made a profit for that game/console
so for PSP and 360 I made a nice profit, but for DS is lost a bit, and failed on both Wii and PS3.
I think ioi is now implementing a multiplier of 1.1 or something, which means you can in fact be slightly less accurate than average and still make money, so I would have been profitable on the DS too (though not much)
Edit: ignore the italics because I was using the accuracy of the average prediction, instead of the average accuracy
-------
If you want to work out the accuracy before the tables are updated simply get the difference between your prediction (or the average prediction) and the actual numbe, take it away from the actual, then divide by the actual and multiply by 100.
So for the Wii:
115,000 - 97,723 = 17,277
97,723 - 17,277 = 80,466
80,446/97,723 = 0.8232043 *100 = 82.32 (2.d.p)
