Based on previous draws, I fitted a linear regression model that can predict the scores for next draws based on number of people selected at each of them. I assumed that in the next draw(s) number of people would be anywhere between 700 to 2500. Based on the assumption, here are some predictions:
#of People Selected Predicted Scores
700 500
750 499
800 497
850 496
900 494
950 492
1000 491
1050 489
1100 488
1150 486
1200 484
1250 483
1300 481
1350 480
1400 478
1450 477
1500 475
2000 459
2200 453
2300 450
2500 443
It seems like to reach score 450, 2300 people needs to be selected. Its a linear regression model and it will perform better and better based on number of data points its been trained on. Right now, there have been 42 draws. So I felt like we have atleast not a bad amount of data to fit a regression model.
#of People Selected Predicted Scores
700 500
750 499
800 497
850 496
900 494
950 492
1000 491
1050 489
1100 488
1150 486
1200 484
1250 483
1300 481
1350 480
1400 478
1450 477
1500 475
2000 459
2200 453
2300 450
2500 443
It seems like to reach score 450, 2300 people needs to be selected. Its a linear regression model and it will perform better and better based on number of data points its been trained on. Right now, there have been 42 draws. So I felt like we have atleast not a bad amount of data to fit a regression model.