Update: The error still persists after 48 hours and carries on. I have updated the numbers.
Update 2: SoloRedCup corrected it and the first column of numbers is the official one. I noticed later that there was an exact factor of 1.3 between both columns. As it turns out – one was a projection and was accidentally copy&pasted.
Just after 6 am in the morning, still in bed I take out my phone to check the progress as I planned to give a continuous commentary about the progress. Instead I’m close to a melt down.
So what is wrong? It is just simple Maths – that’s it!
| Progress last 12 hours | Progress last 12 hours |
| 797297 – 0 = 797297 | 36000000 – 35202703 = 797297 |
| 2384062 – 797297 = 1586765 | 35202703 – 34166106 = 1036597 |
| 3305446 – 2384062 = 921384 | 33457349 – 34166106 = 708757 |
| 4603719 – 3305446 = 1298273 | 32458678 – 33457349 = 998671 |
The progress per 12 hours and the reduction in points needed over the last 12 hours mathematically have to match.
I had planned to make a prediction graph based on Myriad. The idea – the play pattern overall during Aurora should match closely Myriad. I would expect some shifts – but overall there are the same people playing. This is more or less how AI work – just they tend to need a lot more data before they are able to pick up anything.
So here is a different view of the data:
| Myriad | Aurora 1 | Percentage | Aurora 2 | Percentage |
| 1825283 | 797297 | 44% | 797297 | 44% |
| 2284181 | 1586765 | 69% | 1036597 | 45% |
| 1685456 | 921384 | 55% | 708757 | 42% |
| 2315147 | 1298273 | 56% | 998671 | 43% |
| Estimate | ||||
| 29124588 | 16301935 | 56% | 12687402 | 44% |
According to the progress data we started slowly, improved a lot and have been falling back again over the last 12 hours compared to Myriad. If we assume that the points remaining are correct, then we follow very, very closely the data of Myriad.
Both scenarios are possible. My own prediction algorithm that I hoped would be a lot more accurate as the linear extrapolation of Niantic would rely more closely on the second scenario.
It is not inconceivable that play did pick up a lot after the first 12 hours and that the data is much more unpredictable.
The only knowledge we have for 100% is – it can’t be both.
What I just don’t understand: The calculation I do explicitly is just a basic spreadsheet operation – actually one of the most simple ones possible. Take a fixed value (36 million) and take away a value from a column
=36000000 – B2 if you want to calculate points remaining (then copy&paste)
=36000000 – D2 if you want to calculate points achieved (then copy&paste)
You hear a lot of talk about Niantic Calvinball it. In my view this is plain wrong. We never would have overachieved Myriad by such a margin if Niantic would Calvinball it.
But it is these small mistakes in communication and lack of focus on detail for such an event that open you up to these false accusations.
So please Niantic – I hope someone there reads this and corrects it. It might not matter in the wider picture – 13 million or 16 million – we will fail by a large margin. But it will matter for long term trust in you.
Oh – and of course it matters for me. Good luck. I actually enjoy Aurora a lot. Just wrong data is something I can’t stand. Rant over.