I tried calculating what could be the CRS cutoff score for the next draw.
As of April 26, there were 1780 candidates with scores of 450+ (279 +1511). In the May 1st draw, the cutoff score was 450. Let us say that 400 (my guess is this is a conservative estimate) candidates with the score of 450 were selected in that draw, that leaves us with (3350-1780-400=) 1170 candidates. In April 17th draw, the cutoff score was 451, and assume that 100 candidates with the score of 451 were left out in the pool. So, after April 17 draw and April 26, 1680 candidates with scores of above 450 were added to the pool. Between April 26 and May 1, 1170 candidates were added to the pool in 5 days compared 1680 candidates in the prior ten days. By looking at the previous draws, the trend seems to be there is a rush for people to add their ITA profile as the draw date nears. So, we cannot assume that the number of people added to a draw is uniform.
As of May 10, there were 1985 candidates with scores of 450+ (317 +1668). On May 24, there were 5353 candidates (729 + 4624) with the same range of score. We know that there is a more significant influx of profiles created near the draw date. If we subtract the number from May 24 to May 10, we can know how many profiles were created in a cycle. By a cycle, I mean the number of profiles created from the end of a daw to the beginning of the next draw. From May 10th to May 24, there were (5353-1968) 3368 new profile. I rounded this number to 3350 as rest of my calculation were simple to make it the same as the number of people invited in a draw. Note that for every cycle, 3350 candidates are being added, if this trend continues, I don't see a future where cutoff drops below 451 if the draw size is not increased.
In the last draw, there would have been 6700 candidates with scores of 451+ as I believe that 3350 profiles with that score are being added every cycle. At a minimum, there could have been 6550 or 6600 profiles with that score. Of that 6700 candidates, 3350 or half of the pool had scores of 470+. I made a big assumption that profiles are uniformly distributed across the score range of 451-470, that is 3350 candidates in that score range are evenly distributed each score (451, 452) has (3350/(470-450))167.5 profiles. I understand this is an unrealistic assumption and that there would be more profiles towards the lower end of scores. But, this assumption leads to a conservative estimate (that is cutoff score should be lower than what is being estimated, however after a certain point, the drop in cutoff score would be small as there are more backlog profiles in the lower end of score). By making the same assumption, each score will have (167.50/2) 84 new profiles during every cycle.
After the above assumption, the calculation was simple that I need to look at how much backlog is being cleared and how many profiles are being added for every draw.
If a b2b draw happens today on June 5th,
Cutoff today - 458
June 19th - 455
July 3rd - 454
July 17th - 453
July 31st- 452
August 14th - 452
If the next draw is on June 12th,
June 12th - 463
June 26th - 459
July 10th - 457
July 24th - 456
August 7 - 455
August 21 - 454
Assumptions made
1. This is more of an educated guess that I am confident to be accurate, there is a greater ratio of people adding their profiles near the draw date
2. The trend we saw in recent draws with lots of profiles with scores of 451+ is not an anomaly and is not seasonal, as in, we usually have higher scores during this period of the year
3. There is not going to be an upward tick in the number of profiles with scores of 451+
4. The number of profiles in the range 451-470 is evenly distributed, but this assumption should lead to a conservative estimate.
I wish CIC released much more granular information with the previous draw by having a bucket for 451-460 and 461-470. I am waiting for the next set of numbers to be released on how to tweak these numbers.