• Hello Members, This forums is for DV lottery visas only. For other immigration related questions, please go to our forums home page, find the related forum and post it there.

Immigrant NumberVisa Issuances in the Diversity Immigrant

As in 2013? Are you telling they already did it this year?
Yes, on Octobe 22nd 2012. The highest number for Europe during the firs batch was about 31000, and after the second batch is is about 32000. That is about 3% increase to what was open first.

By mail or internet?
Internet.

Sorry but it does not add up.
Why not?
Also if they select only 100k and then remove duplicates how come there is AF106xxx and AS25xxx all at the same time?
After they remove the duplicates they have about 100K. Not before. So, a lot of numbers from high-fraud countires are not used (crossed out during additional checks)



I think they disqualify duplicates right away,

Right after assigne the numbers, they disqualify duplicates.

then randomize the rank order,
They are random from the very beginning.
then start selecting randomly and assign case numbers and then doing some more filtering and keep going until they hit their target.
Right after they assign the numbers, they disqualify duplicates and more (like entries with a non-human face). And then they publish visa bulletin with results, and then release about 50% of winners.
 
Yes, on Octobe 22nd 2012. The highest number for Europe during the firs batch was about 31000, and after the second batch is is about 32000. That is about 3% increase to what was open first.
Can you tell us where you got this from? And by the way, it should be 33088.

They are random from the very beginning.
No. First, numbers are given sequentially in the order of entry submission. Then rank order is randomized, so even if you are the very first one to submit your DV entry, you may end up number 8 million (this part is what they messed up in 2012). Then random selection is done. You need to do some reading ;)

After that is a bit fuzzy. For example there are 52080 selected for Africa, and that is already 52K out of 105K, so they will probably only issue about 25-26K visas to Africa, yet there is a case number such as AF106xxx. So they must have given the case numbers after the selection (otherwise we could've easily seen AF4million) but before filtering duplicates, frauds, double winners etc. (otherwise we couldn't see numbers higher than AF52080). So they must have selected way more than 100K for Africa, and then go through them in rank order to weed out disqualified ones until they hit 52080th selectee whatever his/her case number is. And this is just for Africa.

Anyways, I'd like to see some law articles and regulations rather than hearsay and guesswork. Anybody's guess is as good as mine, but probably not better ;)
 
Can you tell us where you got this from? And by the way, it should be 33088.


No. First, numbers are given sequentially in the order of entry submission. Then rank order is randomized, so even if you are the very first one to submit your DV entry, you may end up number 8 million (this part is what they messed up in 2012). Then random selection is done. You need to do some reading ;)

Totally agree!
About second draw. There was one in October which was confirmed by KKC. I know at least two people that weren't selected in May but in October. Also on this forum there were couple.
 
Can you tell us where you got this from?
Just following another forum where some winners report their wins with dates and rank numbers. That is a Russian forum. Do you read Russian?
I'll answer the second question a little later.
 
I think they disqualify duplicates right away, then randomize the rank order, then start selecting randomly and assign case numbers and then doing some more filtering and keep going until they hit their target
This is incorrect.
They are random from the very beginning.
No. First, numbers are given sequentially in the order of entry submission. Then rank order is randomized, so even if you are the very first one to submit your DV entry, you may end up number 8 million (this part is what they messed up in 2012). Then random selection is done.
First, they cannot disqualify duplicates right away. That is technically impossible. That is a very costly operation, and cannot be done with all entries. Only entries that are selected are checked for duplicates. Check the technical characteristics for the software they use.
Second, entries are kept in the database almost in the order of submission, with small number of exceptions.
Then entries are randomly reordered. And then numbers up to certain number are selected. And then numbers until some number are published on private pages of winners on DOS website. So, in your initial statement you stated "start selecting randomly". That is wrong, they just start selecting. Randomly makes no sense here, because randomization was done before that. So, just a selection is done, not a random selection is done.
My answer "from the very beginning" meant "before the selection".

I'd like to see some law articles and regulations rather than hearsay and guesswork. Anybody's guess is as good as mine, but probably not better
There is Kirit Amin's declaration to the court, as well as supplemental declaration. Both are taken under oath and under penalty of purjury, that is much more than a regular DOS document.

DECLARATION OF KIRIT AMIN

I, Kirit Amin, pursuant to 28 U.S.C. § 1746 do hereby declare the following under penalty of perjury:

1. My name is Kirit Amin. I am employed by the US Department of State where I worked since June 2007. I am assigned to the DOS Bureau of Consular Affairs, where I am the CIO to Consular Affairs and Director of Consular Systems and Technology. In that capacity I oversee all systems related operations involving Consular Affairs, including the computer database and selection program used in the 2012 diversity lottery.

2. As a result of my job position, I have become thoroughly familiar with the computer hardware, software and databases that the State Department relies upon to administer the DV Lottery program. I am also familiar with the process by which the software that the State Department used to conduct the DV lottery was written, installed and implemented.

3. As a preliminary step in this yerar’s DV lottery, the State Department operated a website at which aliens seeking a diversity visa could submit petition during a submission period. The submission period this year began on October 5, 2010 and ended on November 3, 2010 (the “Submission period”). Each petition server as a “lottery ticket” in the DV lottery.

4. After the submission period closed but prior to the DV Lottery selection process the State Department sorted the petitions into different world regions and numbered them in a database according to the order that they were received.

5. Because of the way our database’s internal storage optimization algorithms work, the database moved the physical database location of some petitions that were submitted later in the Submission period, totaling about two percent of the total number of entries, so that they were adjacent to records that had been submitted in the first two days of the Submission period. This database optimization occured prior to the Lottery selection and is a standard data storage protocol used by Oracle brand database software. It had nothing to do with the fact that this particular database contained DV Lottery entries.

6. This year, the State Department used a new computer program intended to randomly renumber the DV Lottery petitions.

7. The programmer who wrote the program, however, made an error that essentially rendered the program ineffective. Instead of instructing the computer to renumber the petitions from entry date order to random order as required by 22 CFR § 42.33(c), the computer program simply selected entries in the existing order which was the order in which they entered plus two percent of applicants reordered as part of resulting from database optimization. Thus, the computer program designed to make selection random failed entirely.

8. I am familiar with regulations of 22 CFR § 42.33(c) that requires that the DV entries be “rank ordered at random be a computer using computer software for that purpose.” In computer software the use of term “random” ordering of a list is a term of art that requires specialized software that generates numbers that are mathematically proven to be random. Not only did the software we used fail to randomize the DV entries here, but the two percent of entries who were at the top of list due to database optimization also fails to meet the definition of random in the regulations because no computer software designed for the purpose of randomizing was used.

9. The computer programmer’s error explains why 98% of the lottery “winners” came from October 5 and 6, 2010, with the remaining two percent of the “winners” submitted on other days in the submission period.

10. The State Department made the results of the selection available on its website on May 1, 2011, without realizing that the programmer’s error had failed to randomize the petitions.

11. During the period in May when the erroneous results were posted, 1 940 615 applicants logged on and checked the results. Of these applicants, 22 316 were notified in error that they had been selected.

I declare, under the penalty of perjury, that the foregoing is true and correct. Executed on July 6, 2011.

/s/ Kirit Amin
Supplemental declaration of Kirit Amin

I, Kirit Amin, pursuant to 28 USC § 1746, do hereby declare the following under penalty of perjury:

1. My name is Kirit Amin. I executed a Declaraion in this action on July 6, 2011 (ECF No 7-1).

2. I submit this Supplemental Declaration to address questions raised by the Court during its July 12, 2011 hearing concerning the 2012 DV Lottery and to provide additional information about the State Department's efforts to comply with the procedures mandated by 22 CFR 42.33(c).

3. As a preliminary step in the DV Lottery the State Department created a website to which aliens seeking diversity visa could submit their petitions during a submission period. The submission period for this year's DV Lottery began on October 5, 2010 and ended on November 3, 2010.

4. As DV Lottery applicants submitted their petitions, a database program captured and recorded the petitions onto a series of hard drives. The database program stored each petition in a physical location on the hard drive, for the most part in the order in which they were received. However due to the database program and the "storage optimization algorithms" there are some exceptions to the order by which database stored petitions, as I discussed in my July 6 Declaration.

5. Because DV Lottery participants submitted petitions at such a high volume, the database program was not always able to store an incoming petition at the physical location on the hard drive immediately adjacent to the location where it had recorded the immediately preceding petitions. When that happened, to keep up with such high volumes and to perform optimally, the database program would record the petitions in a distant location on the hard drives and leave, temporarily, an empty spot or gap on the hard drives adjacent to where it had recorded the immediately preceding petition.

6. This temporary, fleeting inability of the database program to access a particular physical location on the hard drives and store petitions in sequence explains why some entries submitted on October 5 and 6, 2010, were not stored in a physical location on the hard drives alongside the other entries submitted on October 5 and 6. This temporary inability to store records in a particular physical spot on the hard drives also resulted in gaps on the hard drives in the location where the database program recorded nearly all - but not all - of the petitions submitted on October 5 and 6. As the Submission period progressed, the database program filled these gaps in the hard drives with petitions submitted later in time. This backfill process - which I referred to in my July 6 declaration as "internal storage optimization algorithms" (Decl. § 5) - explains how and why the database program recorded the two percent of the petitions submitted after October 5 and 6, 2010, in the sections of the hard drives containing nearly all the petitions submitted on October 5 and 6 2010. The database program was designed to minimize the number of these gaps on the hard drives so it could later access the data more quickly and efficiently.

7. After the database program recorded each petition submitted over the Submission period, the database program rank-ordered each petition in the order they were located on the physical hard drives which, as I noted earlier, generally (but not exclusively) corresponded to the date on which they were submitted. Also, as required by the DV Lottery program, the database program applied region and country limitations to the selection to ensure no single region or country dominated the selection. These limitations, along with the manner in which the database program stored the petitions on the hard drives, account for the fact that some petitions entered on October 5 and 6 were not selected: there were enough petitions with a lower physical location in the database to satisfy the region and country limitations.

8. At the conclusion of this data storage process (step one), the State Department must initiate a second process in which it rand-orders the petitions again, but this time at random, using computer software designed for this purpose (the "Randomizer program").

9. Next, the State Department must initiate a third process in which it selects petitions from each world region in their rank order as determined by the Randomizer program as winners of the DV Lottery.

10. This year, the State Department used a new computer program intended to server as the Randomizer program.

11. The programmer who wrote the Randomizer program, however, made an error that rendered the Randomizer program ineffective. Instead of instructing the computer to select DV Lottery winners based on the rank ordering of the Randomizer Program in step two, the computer program simply selected entries in the order in which the database program stored petitions on the hard drives in step one. Thus, the Randomizer program, which was designed to make the selection random, failed entirely to achieve that goal.

12. I understand the Court has inquired about the State Department's interpretation and definition of "random" that appears in relevant DV Lottery statutes and regulations, including 22 CFR §42.33(c), which requires that the DV Lottery petitions be "... rank ordered at random by a computer using computer software for that purpose." In computer software, a "random" sequence is one in which the numbers in the sequence are generated as if they were independent draws from a well-mixed vessel where each number is represented once in the vessel. This random process embodies qualities of unpredictability and equal probability. In other words, the "random" rank-ordering of a list is a term of art that requires specialized software that generates numbers that are mathematically proven to be without any definite aim, direction, rule or method.

13. The first step required by 22 CFR §42.33(c) - the process executed by the database program - is not designed to produce random results when the database program recorded each petition in a physical location on the hard drives, it did so with a definite aim: the aim to fill the hard drives in consecutive order to the greatest extent possible while also optimizing data storage and retrieval and minimizing the number of empty spaces on the drives that were interspersed with portions containing data. The database program also recorded each petition with direction: it started recording petitions in the first available physical location on the hard drives and recorded each additional petitions, to the greatest extent possible, in a location on the hard drives physically adjacent to the space where it recorded the preceding petition. Finally, the database program recorded the petitions with a rule or method designed to optimize data storage and retrieval and to minimize the empty portions of the drives that were interspersed with portions containing data.

I declare under penalty of perjury that the foregoind is true and correct.

/s/ Amin.

More answers ahead.
 
Last edited by a moderator:
After that is a bit fuzzy. For example there are 52080 selected for Africa,

So, we are talking about DV-2013

and that is already 52K out of 105K
Out of 105,625, what would be rounded to 106K

, so they will probably only issue about 25-26K visas to Africa,
About that.

yet there is a case number such as AF106xxx.
That is for numbers which are already open. There are some hidden numbers among those 52,080. They have higher numbers, up to about AF165xxx

So they must have given the case numbers after the selection (otherwise we could've easily seen AF4million)
No, numbers are given before the selction, during randomization phase.

but before filtering duplicates, frauds, double winners etc.

Correct

(otherwise we couldn't see numbers higher than AF52080).
As Wolfsdorf mentioned, they assign numbers like AF00001, AF00002, e t.c. for all entries. Then they disqualify junk, and select numbers up to AF165xxx (totally 52,080 numbers, the rest 165,000-52,080=113,000 are junk and are disqualified). And then they release numbers up to AF106xxx (about 52,080*106,000/165,000=33,500 entries out of 52,080) and keep the rest of numbers hidden. In case of necessity they release additional numbers (they still have about 52,080-33,500=18,580 more numbers in the original set of 52,080 that are hidden up to that point), like what they did on October 22nd for DV-2013.

So they must have selected way more than 100K for Africa, and then go through them in rank order to weed out disqualified ones until they hit 52080th selectee whatever his/her case number is. And this is just for Africa.
Right. And those 52,080 entries are reached when the number of about AF165xxx is reached.
 
Last edited by a moderator:
The Case Of The Case Number

Hi raevsky, thanks for the useful information!

Some of my thoughts

First, about Mr Amin testimony. To put in a nut shell, Amin defends the decision to nullify the first draw of DV12 and explain why it was not random as the regulations dictate.

The programmer probably did something like "SELECT * FROM ENTRIES" to reorder the entries - but this just returned the entries approximately in the order they were entered to the database... This is why QA is so important...



There are some points raised in this thread which I would like to challenge.

So that the case numbers will look like AF00000001, AF00000002, etc.

The regulation says that "the first entry randomly selected will be the first case registered; the second entry selected will be the second case registered etc.". This does not mean that the case numbers are consecutive e.g. 1,2,3,4,5. The second case registered will be 'second' in order but will not necessarily have the number "2". This is not dictated by the regulations. In fact, as far as I understand the process, the numbers are not consecutive. We can have AF001 then AF003 then AF004 then AF007 and so on.


First, they cannot disqualify duplicates right away. That is technically impossible....Right after they assign the numbers, they disqualify duplicates and more (like entries with a non-human face)

Part of this makes sense and part doesn't.

There is no sense in checking for duplicates after selecting the entries. The whole idea is that they check this before the selection so that participants will not increase their chances by submitting many entries. Checking for duplicates is simple. What identifies an entry is the following: First, middle & last name, date and place of birth. They assume that no two persons have those same features. If I submit two entries but change one of the features above, and that entry is selected, I will fail to provide evidence of my identity in the interview (unless I forge my birth certificate or passport).

Checking the photo is a task for a human, I agree. Checking other requirements is done during the interview.

So my understanding is that duplicate entries cannot explain the gaps between case numbers.

An invalid photo also cannot explain why we have case number 73,398 in DV12 Asia while there are only 15K selectees (or EU28XXX in DV13 etc).

However, I do see a direct relation between the high case numbers for Asia, for example, and the selectees/participants ratio of Bangladesh over the years. The lower this ratio was, the highest the case numbers for Asia were. We cannot assume that there are many 'non human' photos from Bangladesh because in DV12 only about 2k were selected from there, yet Asia case numbers were extremely high! In DV13 Bangladesh was not in the game and Asia case numbers are very low (I would say 'normal').

There should be another explanation for those gaps between case numbers. the 'disqualified entries' explanation cannot be it.
 
The regulation says that "the first entry randomly selected will be the first case registered; the second entry selected will be the second case registered etc.". This does not mean that the case numbers are consecutive e.g. 1,2,3,4,5. The second case registered will be 'second' in order but will not necessarily have the number "2". This is not dictated by the regulations. In fact, as far as I understand the process, the numbers are not consecutive. We can have AF001 then AF003 then AF004 then AF007 and so on.
I totally disagree with you and agree with former President of American Lawyers Immigration Association Wolfsdorf. Numbers are consecutive with junk excluded afterwards. That is simpliest in relization, does no artificial numbering and leads to the right figures.
In our example, among 165000 selected entries from Africa 113000 are complete junk, and only 52080 non-junk numbers are the final selection.


There is no sense in checking for duplicates after selecting the entries.
This is an expensive operation and it makes sence to decrease the amount of checks about 100 times by checink not all 20 million entries versus 20 million entries, but only 100,000 ventries versus 20,000,000 the rest.

The whole idea is that they check this before the selection so that participants will not increase their chances by submitting many entries. Checking for duplicates is simple.
It is the most difficult thing in the whole algorithm.
What identifies an entry is the following: First, middle & last name, date and place of birth.
No. Date of birth and the photo.

They assume that no two persons have those same features.
No two persons have the same picture and DOB.

If I submit two entries but change one of the features above, and that entry is selected, I will fail to provide evidence of my identity in the interview (unless I forge my birth certificate or passport).
Name is easy to change, especially in countries with non-Roman alphabet. But you cannot change picture and DOB.


Checking the photo is a task for a human, I agree.
Totally wrong. It is done by computer. Humans are needed to double check the computer determination. For those 100,000 winning entries computer indetifies closest pictures among 10,000,000 and human double checks. Name and place of birth are not used, easy to forge.

Checking other requirements is done during the interview
Consul verifies that picture corresponds to the person. And checks date of birth in the documents

So my understanding is that duplicate entries cannot explain the gaps between case numbers.
Their amount is rather low. See www.gao.gov/new.items/d071174.pdf

An invalid photo also cannot explain why we have case number 73,398 in DV12 Asia while there are only 15K selectees (or EU28XXX in DV13 etc).
74,000 - this is a maximal open number. There are much more, totally about 141K including hidden numbers from Asia.


However, I do see a direct relation between the high case numbers for Asia, for example, and the selectees/participants ratio of Bangladesh over the years.
Me too

The lower this ratio was, the highest the case numbers for Asia were. We cannot assume that there are many 'non human' photos from Bangladesh because in DV12 only about 2k were selected from there,
About 99.7% of Bangladeshi entries are complete junk, entries with invalid photos, when agents enter white pages books into the database, with dummy pictures. That was quoted in congressional testimony.

yet Asia case numbers were extremely high! In DV13 Bangladesh was not in the game and Asia case numbers are very low (I would say 'normal').

There should be another explanation for those gaps between case numbers. the 'disqualified entries' explanation cannot be it. .
It is.
 
Yes, on Octobe 22nd 2012. The highest number for Europe during the firs batch was about 31000, and after the second batch is is about 32000. That is about 3% increase to what was open first.

Internet.

Why not?

After they remove the duplicates they have about 100K. Not before. So, a lot of numbers from high-fraud countires are not used (crossed out during additional checks)





Right after assigne the numbers, they disqualify duplicates.


They are random from the very beginning.
Right after they assign the numbers, they disqualify duplicates and more (like entries with a non-human face). And then they publish visa bulletin with results, and then release about 50% of winners.

actually it was less than 31K I am 29 *** and was selected october 22nd. I think in may were notified around 25K-27K and then in october from that number up to 33K or 33088 as is the total number of winners from europe
 
I checked it. I see 27xxx notified in May and 28xxx-30xxx on October 22nd
Where did you see the number 33088?
If so, that is quite a lot of numbers in the second batch from Europe. Not huje, but not tiny as wellBut there are about 61K total numbers from Europe. About 33K are open right now.
 
Last edited by a moderator:
I don't know for a fact but what he said makes sense. Some people may have not been aware of the second draw so they didn't apply for a visa even though they have been selected. Also, due to the lawsuit, DoS may not have done the extra selection in case they are not filling the 50K quota. Or for some odd reason a lot more people were disqualified in 2012 but that is not very likely. At any rate it shouldn't matter for 2013.

Are you asking if there is going to be another lawsuit this year? Up to this moment, there was no reason for it published.
Or do you want to know if there is going to be a different problem this year leading to the same result? Up to this moment, there was no reason for it published.
 
I just cant believe that the main reason why only 38k visa has been issued on 2012 is that less people have been responding for their selection. Well they will be always hungry people for those visa !
 
I checked it. I see 27xxx notified in May and 28xxx-30xxx on October 22nd
Where did you see the number 33088?
If so, that is quite a lot of numbers in the second batch from Europe. Not huje, but not tiny as wellBut there are about 61K total numbers from Europe. About 33K are open right now.

Girl from my country notified in October has CN 25xxx.
 
I just cant believe that the main reason why only 38k visa has been issued on 2012 is that less people have been responding for their selection. Well they will be always hungry people for those visa !

No, that is not the main reason. The main reason was that there was no second batch. And that was because of the court.
 
Girl from my country notified in October has CN 25xxx.
EU number?
What is the last date she checked her status and there was no win?
What is the first date she saw the win?

Previous second batches never happened this way. Only numbers from larger diaposon were released during second batches. Most likely, she just did not check her status before.
Also, in order to check the status correctly you need to enter information correctly. If you enter it incorrectly, you might not know that is was enterer in a wrong way. And could think that you were not selected.
They changed that recently, but I am not 100% sure when exactly

I am certain if her number is EU25xxx, it was shown already in May. If she did not see it, she did something wrongly.
 
Last edited by a moderator:
raevsky,

Thanks for the testimony info. I was familiar with some of it but didn't read the whole thing. These people give contradicting information; in some press conference a DoS official stated that there are two randomization steps. And this Amin dude says that they eliminate entries w.r.t regional quotas BEFORE they randomize the rank order ?!!

Another thing is, you are telling us that you know the "closed" numbers? Well I'll leave it at that and won't even ask how. Have a good day :)
 
they eliminate entries w.r.t regional quotas BEFORE they randomize the rank order
What exactly do you mean? Which statement of his do you refer to?

in some press conference a DoS official stated that there are two randomization steps
The young girl from visa office? Are you talking about http://fpc.state.gov/198409.htm ?
King does not know anything, Thurmond does. Ignore everything King says, consider only what Thurmond says.
Those steps are not in their documents, that is part of computer progem, and it is not easy to isolate one step from another.
you are telling us that you know the "closed" numbers
No, I do not know. But I have a way to estimate them.
Well I'll leave it at that and won't even ask how
That is up to you, but I have an explanation for everything. Everything comes from public data.
 
Last edited by a moderator:
Actually, DV-2012 revealed a lot of things to me, because of the court proceeedings and because of the bug itself. Also, because I closely followed the story with Uzbekistan.
 
This is from Edson Congressional Testiminy (about Bangladesh):

Nor is this sort of abuse limited to applicants who choose to enter the diversity visa lottery. Consular officers have many times seen unscrupulous agents enroll large groups of people, including in Bangladesh the phone book, so that they could then extort money from the legitimate applicants who didn't in fact apply, or from other people that they sell the winning slot to as they apply for the visa
 
Top