Do you think they catched duplicate entries?

[tmfl]

That means that those who are registered and those who are notified count to 100,000 together.

X is for A and X' is for B. It could be the case, although that's not simple and sounds ad-hoc. The other interpretation I mentioned is simpler and also it could be. Not clear.

More is here because typically it is about 106000-110000. Not because it is 200000 with dependents.
And less than half means that for more than 100,000 family members there are only 50,000 visas. Less than half of more than 100,000

But sometimes less than 100,000 winners in fact. Your explanation sounds ad-hoc. 200,000 with derivatives is always more than 100,000 and always less than half.

Not clear....

 

[raevsky]

4K out of 82K winners in DV2007, 5% of selectees, suggesting 5% of total entrants were considered illegitimate, that's it.
If 82K includes derivatives, it should be less than 5% because 4K is number of entries (PA only).
And, at this point, perhaps KCC recognizes some other kind of fraud, such as phonebook fraud, but there is no evidence that they detected & disqualified them.

What about this?

http://www.state.gov/s/d/rm/rls/perfrpt/2005/html/56332.htm

In 2005, more than 7.5% of winning entries were disqualified as duplicate entries using facial recognition. Over 100,000 more duplicate entries were eliminated using facial recognition and matching tech nologies before the winners were selected.

7.5% - it is 7,622 (from the previous quote for 2005). That means that the total amount was about 100,000. Additionally, over 100,000 - 7622 = over 92378 were eliminated using matching tech nologies, without using facial recognition software. Totally about 50% were eliminated before the winners were selected. Matching tech nologies sounds like comparing files for direct matching.

Those 86% from Bangladesh and Nigeria are eliminated using matching tech nologies. At this point about 2/3rds of all entries worldwide submitted are junk which is eliminated by matching tech nologies. As you see, it was about 50% in 2005.

 

[raevsky]

Also, in DV-2003

http://www.state.gov/documents/organization/53112.pdf

Used FR technology to disqualify over 20,000 from the annual DV lottery for filing duplicate entries. To evaluate FR’s full potential for combating visa and passport fraud, launched a facial recognition pilot for nonimmigrant visas (NIV)

And yes, I do not see any direct evidence that 100,000 - that is including dependents. However, from my point of view the easiest way to interpret those statements is to assume 100,000 include dependents

 

[tmfl]

7.5% - it is 7,622 (from the previous quote for 2005). That means that the total amount was about 100,000. Additionally, over 100,000 - 7622 = over 92378 were eliminated using matching tech nologies, without using facial recognition software. Totally about 50% were eliminated before the winners were selected. Matching tech nologies sounds like comparing files for direct matching.

From what I cited in my previous post, matching procedure applied to all entries before selection process, then facial recognition software applied to selectees. So, in my understanding 92378 out of 9.5million means approx. 1%, plus 7622, 7.5% of selectees suggesting 7.5% of all entries , totally 8.5% of applicants were supposed to have fraudulent information, or I may be wrong...

 

[tmfl]

It is clearly defined as 6 months. Definitely, 1-year photo is beyond any possible limits.

I tried, but failed to get this. Could you please let me know where it is?

 

[raevsky]

The entries are screened electronically for exact duplicate digital photographs or biographic information;
exact duplicates are disqualified.
Next, all principal applicant entries selected in the lottery are checked with facial recognition software against galleries (by region and by gender) drawn from the entire e-DV database to further eliminate duplicate entries based on photo matches.

From what I cited in my previous post, matching procedure applied to all entries before selection process,

That is not so obvious, though it would be theoretically technically possible, it does not make a lot of sense.
I would give another ad-hoc explanation. They produce a set of randomly selected entries, and then they start matching them against the whole database. It makes no sense to match the whole database vesus itself anyway. Also, it would be completely unbeleivable that 7.5% are identified by FR and only 1% by matching procedures, provided extremely high fraud in several countries. Though it was much more rare in 2005. It would be much more likely that was 50% in DV-2005 (and more close to 60%-65% in DV-2011)

Now about 6 months. That is from photo instructions. That requirement is relatively new, it did not exist in 2003, and not even in 2005, though I dod not remember when exactly it was put in place.

The image must be from a recent (within 6 months) photo of the entrant.

I put a quote from DV-2011. The same is true in DV-2012.

As we see here on the forum requirements like that one are not well known, and that is why could produce enormous percentage of disqualifications. Anyway, it cannot be just 1%, 65% is much more close according to the situation on the forum

 

[tmfl]

That is not so obvious, though it would be theoretically technically possible, it does not make a lot of sense.
I would give another ad-hoc explanation. They produce a set of randomly selected entries, and then they start matching them against the whole database. It makes no sense to match the whole database vesus itself anyway.

I agree, in the meantime,,,

Also, it would be completely unbeleivable that 7.5% are identified by FR and only 1% by matching procedures, provided extremely high fraud in several countries. Though it was much more rare in 2005. It would be much more likely that was 50% in DV-2005 (and more close to 60%-65% in DV-2011)

Again, despite high fraud level, duplicates are decreasing across year, regardless of interpretation (8.5% or 50% or whatever). I speculate multiple entries which do not look alike (each looks like unique) and/or other types of fraud are/is increasing. So far, I don't see any official/reliable source in which they say they have successfully removed plenty of those types of entries during selection process.

Now about 6 months.

I got it, thanks so much. I did not know that. DOS could run some screening comparing to previous database. Again I don't know of any documents suggesting they are doing, but they could.

 

[tmfl]

And yes, I do not see any direct evidence that 100,000 - that is including dependents. However, from my point of view the easiest way to interpret those statements is to assume 100,000 include dependents

I don't disagree with your assumption. Quite frankly I don't know which is correct. I always pay attention to this issue whenever I read DV-related documents, and so far, either could. I check both cases when I calculate wining probability or other stuff. For example, going back to FY1996 table I quoted before, more than 2X of winners of Nigeria (15K vs 7K winners) showed up at the interview and some countries were slightly more than winners, others were less than winners. Why more than 2X winners? Winners do not include dependents with some add-on children? Winners do include dependents with many "add-on children" and "pop-up" spouses? I don't know
This table also explains an example for variability of "demand" across country. It is variable.

 

[raevsky]

I speculate multiple entries which look like different each other (each looks like unique) and/or other types of fraud are/is increasing

What do you mean? Give an example. If those entries are not caught, they are irrelevant to explain those 86%. Ig they are caught, how would they differ from other types of fraud?

Again I don't know of any documents suggesting they are doing, but they could.

This is a recent rule. There was a reason it was put in place. What could this reason be?

 

[tmfl]

What do you mean? Give an example.

That's not my business. I am not a fraudulent person and I am not in charge of fraud detection. Duplicates are decreasing, still high fraud level, no evidence of improvement of visa issuance (potentially even worse), taking all these into consideration other frauds which are not detectable could be increasing.

If those entries are not caught, they are irrelevant to explain those 86%.

This circulating logic is also not of my business.

What could this reason be?

I appreciate it if you could explain with reliable source, thank you.

 

[raevsky]

Winners do not include dependents with some add-on children? Winners do include dependents with many "add-on children" and "pop-up" spouses? I don't know

I do not know either.

Officers in Accra estimated that up to 50 percent of DV adjudications there involve a “pop-up” spouse; officers in Addis Ababa said that the percentage of sham DV marriages at their post could be as high as 90 percent and a cable from the post stated that DV-related marriage fraud was “rampant.”

KCC employees found 804 “pop-up” spouses or children; applicants from Ghana, Nigeria, and Ukraine had more “pop-ups” than any other country.

So, 2x is possible just using po-ups.

I am not a fraudulent person and I am not in charge of fraud detection

Who are you? What are you doing on this forum? It looks like you did not apply for DV for yourself, otherwise you would know about 6 months rule. I would hardly beleive you could miss it once you read it once.
You understand logic and have at least basic ideas about statistics. Are you a technical person who works for a lawyer? Are you a lawyer with engineering/math background?

I was a DV2010 winner

Or if that is true, does it mean 6 month rule was not in place in DV-2010?

I appreciate it if you could explain with reliable source, thank you.

No, I do not have any reliable source. I could only guess. I speculate that is specifically an anti-fraud measure, not necessarily of immediate use.

 

[tmfl]

Who are you? What are you doing on this forum?.

Why do you need such privacy information? For your bias?

Or if that is true, does it mean 6 month rule was not in place in DV-2010?

That is true, and it does not mean like that. More possible stories

 

[raevsky]

Why do you need such privacy information? For your bias?

No, just interested. I do not have any bias.

More possible stories

What are other possibilities?

 

[tmfl]

No, just interested. I do not have any bias.
What are other possibilities?

A lawyer in a small firm, who took care of my H1B visa, did my DV entry on my behalf for free. I sent him a new photo every year anyway. The firm was about to go bankrupt by the time I won, so I did further processing on my own. That firm did not ask me to pay actually. After wining, I learned, learned, and learned a lot and collected information as much as possible. Eventually I got GC. I started learning how DV works after I was selected.

I am doing 2 professional tracks and one is PhD in neuroscience which has been my second priority. I am not a statistician, but I am familiar with basic statistical analysis, otherwise I couldn't publish 30+ research papers.

I am here because I want to collect info for my friend playing DV 2012. I intended to leave this forum as soon as I confirmed Entrant Status Check Site works properly, however however,,,,

 

[raevsky]

What is your first priority?

 

[bentlebee]

I do think they will throw out duplicates....JMO. Maybe they don't catch them all in the beginning but when you win and you have entered twice you might not get the NL and you are out of luck. In life you get what you deserve!

 

[al_lupoo]

Really? Half of the Turkey is geographically in Europe, and DOS beleives Turkey is chargeable under Europe.

HALF???? please check The European portion of Turkey, contains 3% of the nation's surface

 

[raevsky]

And yes, I do not see any direct evidence that 100,000 - that is including dependents. However, from my point of view the easiest way to interpret those statements is to assume 100,000 include dependents

I have another indirect proof that is with dependents.
If winners do not include dependents, then it is likely the number of visas issued will be larger for some countries than the amount of winners.
If winners do include dependents, that will never happen.

That never happens for DV-2011 for countries with decent amount of winners (more than 4), but comes rather close for Nepal (2132 winners and 1936 visas, 91%).
On the other hand, that happens with Greece in previous lotteries, but level for fraud for Greece (about 17% of entries are thrown out) is much higher than for Nepal (about 2%), other types of fraud could be higher for Greece too, and I assume that was because of pop-ups and they just found the way to combat it for 2011.

 

[raevsky]

HALF???? please check The European portion of Turkey, contains 3% of the nation's surface

OK. I was wrong. Much less than half. However, there are a lot of countries that are located purely in Asia, but which are still charged to Europe. Like Uzbekistan, Turkmenistan, Tajikistan and some other post-Soviet states.

 

[tmfl]

I have another indirect proof that is with dependents.
If winners do not include dependents, then it is likely the number of visas issued will be larger for some countries than the amount of winners.
If winners do include dependents, that will never happen.

That's what I've done before. Because average probability of visa issuance per winner should be 25% (200K vs 50K visa) or 50% (100K vs 50K visa), even 2 winners in a country could mask what we want to see in this context.
Visa issuance of a country with only one winner is all or nothing.

Following is the cases having more than 100% of issuance. Of course only one winner was reported.
SAINT KITTS AND NEVIS 200% for DV2008 and 2009
BAHRAIN 500% for DV1999
I did not take note though, I found one more, 5 issue/4 winner.

Most likely winners do not include dependents, and it is still likely that winners including dependents with lawful marriage/adopted child could happen between entry and interview period.