my old EU prediction

[raevsky 0]

The number of european CNs above 54000 in DV-14 is far less than 0.1% and is statistically not sound. They could be ignored in any kind of statistical calculations

[raevsky 0]

You missed it.

OK, based on the followin max numbers:

AF 116000

AS 23000

EU 54000

NA based on July number current

OC ~2999

SA ~2999

I got the following final numbers for September:

AF 83000

AS about 15000

EU 38700 max (more likely below 38000)

NA current

OV 1600

SA 1600

Also, I am expecting DV-15 EU max number between 39K and 43.5K, based on about 45500 max EU number.

When I know max DV-15 CNs for other regions (I do not know them yet), I willcome up with other DV-15 numbers.

I still got an error in other than EU calculations.

I would expect max CNs in DV-14 to be:

AF 105000

AS 16100

EU 38000 (not more than 38700)

OC 1750

SA 1650

I still got an error in other than EU calculations.

I would expect max CNs in DV-14 to be:

AF 105000

AS 16100

EU 38000 (not more than 38700)

OC 1750

SA 1650

Max passing CN, not max CN

[vladek15425 7]

You missed it.

OK, based on the followin max numbers:

AF 116000

AS 23000

EU 54000

NA based on July number current

OC ~2999

SA ~2999

I got the following final numbers for September:

AF 83000

AS about 15000

EU 38700 max (more likely below 38000)

NA current

OV 1600

SA 1600

Also, I am expecting DV-15 EU max number between 39K and 43.5K, based on about 45500 max EU number.

When I know max DV-15 CNs for other regions (I do not know them yet), I willcome up with other DV-15 numbers.

Holly cow !

How come AF will end at 83k ! Impossible ?!

Can you elaborate on that ?

[Britsimon 18]

I still got an error in other than EU calculations.

I would expect max CNs in DV-14 to be:

AF 105000

AS 16100

EU 38000 (not more than 38700)

OC 1750

SA 1650

Max passing CN, not max CN

Raevsky, you seem to be ignoring the current data in terms of the currently achieved case numbers and the available Ceac data. For example SA is already close to your September number and the Ceac data is not showing it being near any limits. Any thoughts on that?

[vladek15425 7]

Raevsky, you seem to be ignoring the current data in terms of the currently achieved case numbers and the available Ceac data. For example SA is already close to your September number and the Ceac data is not showing it being near any limits. Any thoughts on that?

Exactly ! Like his prediction for SA how on earth will end up at 1650

If its already at 1550 !

Raevsky didn't get his turkey for breakfast

[raevsky 0]

I would not mind Africa being 105K. Just another typo in calculations. I do not see errors inmy calculations for OC and SA, but I do not have a reliable statistics of CNs. If it is like 10% more than 2999, the resulting value would also be 10% more.

I hope to put my spreadsheet with calculations in a couple of days, after I double check everything. By now since morning I already found a couple of typos for Africa.

Regarding compliance with existing data, I do not care too much about that as well. Europe is the best learned region for me, with maximum amount of statistics, and cutoff is already 36300 for July. That is pretty close to the max I believe in - 38700. That is just 7% more.

1550 + 7% is 1650, and 1300 plus 7% is 1390. My both numbers are not lower.

[rafikbo76 2]

If raevsky talk about AF going to 105K, I definitly should prepare my docs

[MalcomX 1]

thanks for your prediction but EU is already at 36,300 for July VB, how comes that in September we will have only 38700, doesn't make sens for me ...

You missed it.
OK, based on the followin max numbers:
AF 116000
AS 23000
EU 54000
NA based on July number current
OC ~2999
SA ~2999

I got the following final numbers for September:
AF 83000
AS about 15000
EU 38700 max (more likely below 38000)
NA current
OV 1600
SA 1600

Also, I am expecting DV-15 EU max number between 39K and 43.5K, based on about 45500 max EU number.

When I know max DV-15 CNs for other regions (I do not know them yet), I willcome up with other DV-15 numbers.

[MalcomX 1]

lololololol .... as soon as it's not sloner!!!!!!

If raevsky talk about AF going to 105K, I definitly should prepare my docs

[szym 2]

The number of european CNs above 54000 in DV-14 is far less than 0.1% and is statistically not sound. They could be ignored in any kind of statistical calculations

Hm, why is it so, that at the end (highest CNs) we hardly can find a real CN? What is the reason for this?

I know that the density at low CNs is high, and later it decreases because special countries start maxing out. This would mean that at some point (once all special countries reach the end of their CNs) the density should be constant. So why at the end we have less than 0.1% real CNs?

What would you say about CNs in EU DV2015? Where is the point (CN) from which you would not include higher CNs in calculations. In DV2015 for EU we see a lot of 43xxx, some 44xxx and (I think) single numbers above 45xxx. I think there was only one guy from Italy with 47xxx, but I didn't hear about anyone else with 47xxx.

[raevsky 0]

Hm, why is it so, that at the end (highest CNs) we hardly can find a real CN? What is the reason for this?

I know that the density at low CNs is high, and later it decreases because special countries start maxing out. This would mean that at some point (once all special countries reach the end of their CNs) the density should be constant. So why at the end we have less than 0.1% real CNs?

What would you say about CNs in EU DV2015? Where is the point (CN) from which you would not include higher CNs in calculations. In DV2015 for EU we see a lot of 43xxx, some 44xxx and (I think) single numbers above 45xxx. I think there was only one guy from Italy with 47xxx, but I didn't hear about anyone else with 47xxx.

One of possible reasons why cases above 54000 are extremely rare could be because they just do not exists, and are typos or intentional disinformation. Anyway, in one of Russian forums with about 300 cases there is none avove 54000. That is still more than 0.1% though... Even one case out of 300 would be 0.3%.

I do not believe they exist, but even if they do exist that is not a uniform distribution anyway. That is my justification not to consider them.

Regarding 44xxx vs 45xxx CNs in Dv-15 I use German Tank Problem (search on wiki). Even if we do not see 45xxx cases yet, and we see just 44xxx cases, the expected max value is a little higher than that. The more cases total we see, the more expected max value is closer to the max value seen. With 100 cases if max value seen is 44999, then expected max value would be 45500. That is my justification to consider 45xxx number as max CN even though I have never seen numbers above 44xxx.

[raevsky 0]

BTW, In DV-12 I saw EU451xx number, and that was a real number from Ankara consulate schedule website. I initially thought that was a typo, and I email the consulate to confirm. They confirmed that was a real number. I have not seen any other number above 32000 in DV-12. I suspect this number "appeared" during notification process, when KCC tries to notify winners about their status when they lacked DS-230 forms. I think this number was not an open winning number before those notifications started, and it was open as a result of a software bug or so. Anyway, this kind of number could be safely ignored in all type of statistical calculations, just because those numbers are extremely rare.

[szym 2]

One of possible reasons why cases above 54000 are extremely rare could be because they just do not exists, and are typos or intentional disinformation. Anyway, in one of Russian forums with about 300 cases there is none avove 54000. That is still more than 0.1% though... Even one case out of 300 would be 0.3%.

I do not believe they exist, but even if they do exist that is not a uniform distribution anyway. That is my justification not to consider them.

Regarding 44xxx vs 45xxx CNs in Dv-15 I use German Tank Problem (search on wiki). Even if we do not see 45xxx cases yet, and we see just 44xxx cases, the expected max value is a little higher than that. The more cases total we see, the more expected max value is closer to the max value seen. With 100 cases if max value seen is 44999, then expected max value would be 45500. That is my justification to consider 45xxx number as max CN even though I have never seen numbers above 44xxx.

OK. I get it. Thanks for the info.

My number is 2015EU44xxx, so you can count me in to that 44xxx list

When can we expect CEAC data from DV2014 range 32950-36300 to see how many holes are in this range?

[raevsky 0]

Interviews are scheduled around 15th of each month. I do not expect percentage of holes to be different from the previous interval, or even 25400 through 30700.

Another illustration could be this one. provided numbers 47xxx do in fact exist as normal numbers, probability to meet a number between 46000 and 46999 should be the same as numbers between 42000 and 42999. Because we saw quite a number of CNs between 42000 and 42999 and I did not see any between 46000 and 46999, I would assume that is safe to assume for calculation purposes that numbers 46xxx and 47xxx do not exist (or are extremely rare and statistically behave as they do not exist) . So I ignore numbers 46xxx and higher. At the same time I do not ignore numbers 45xxx even though I did not see any of those, just because I saw about 100 CNs at most. So, for instance, if the max number is in fact 45500, it would be normal not to see 45xxx numbers if we have just 100 CNs, but it would be unusual not to see any 45xxx numbers if we have 500 CNs.

[Britsimon 18]

I would not mind Africa being 105K. Just another typo in calculations. I do not see errors inmy calculations for OC and SA, but I do not have a reliable statistics of CNs. If it is like 10% more than 2999, the resulting value would also be 10% more.

I hope to put my spreadsheet with calculations in a couple of days, after I double check everything. By now since morning I already found a couple of typos for Africa.

Regarding compliance with existing data, I do not care too much about that as well. Europe is the best learned region for me, with maximum amount of statistics, and cutoff is already 36300 for July. That is pretty close to the max I believe in - 38700. That is just 7% more.

1550 + 7% is 1650, and 1300 plus 7% is 1390. My both numbers are not lower.

OK Raevsky - I see the rationale. You are looking from a statistical exercise point of view - which is valid of course, even though the resulting numbers may look odd. Oddly, KCC don't seem to be following the program... perhaps they don't understand statistical analysis.