I'm going to try something new
![Cheesy](https://bitcointalk.org/Smileys/default/cheesy.gif)
This topic was created on July 28, 2020, 09:22:31 PM (forum/Roobet time)
That means I need to predict total wagers on July 30, 2020, 09:22:00 PM.
Let's see what happened the week before:
July 21, 2020, 09:22:00 PM: 301,079,265
July 23, 2020, 09:22:00 PM: 307,176,680
Difference: 6097415
Graph:
![Image loading...](https://ip.bitcointalk.org/?u=https%3A%2F%2Floyce.club%2Fother%2Froobet2.gif&t=665&c=IPFQaD_XDBPIxg)
This is not a straight line, and that's probably why I never won using linear extrapolation.
Now let's get to work for my new guess:
July 28, 2020, 09:22:00 PM: 323,171,027
July 30, 2020, 09:22:00 PM:
![Huh](https://bitcointalk.org/Smileys/default/huh.gif)
,???,???
Based on just last week's difference, I'd get to 329,268,442. But my gut feeling tells me there's a difference per week, so let's see how many bets were placed the 2 days before:
July 19, 2020, 09:22:00 PM: 295,638,641
July 21, 2020, 09:22:00 PM: 301,079,265
Difference: 5440624
July 26, 2020, 09:22:00 PM: 316,601,085
July 28, 2020, 09:22:00 PM: 323,171,027
Difference: 6569942
That's 20.757% more than a week earlier! So I'm increasing my estimate by 20.757%: 330,534,082
Rounding down gives my prediction:
LoyceV: 330,534,000
On second thought, I've now included the weekend for my 20.757% calibration, and that's probably not going to work out. In 4 days, I'll have enough data to calibrate using 2 days exactly 2 weeks earlier. That should work better.