Now that DV2018 is over, and the dust is settling, it is natural that people want to know why the cutoff number in EU was so low compared to previous years, what happened, and will that repeat.

So – I have been looking at the data and wanted to give some analysis about that, just about EU which was the single biggest surprise from DV2018.

Generally speaking EU is a region that was previously “largely” predictable from year to year. Success rates, response rates, derivative rates were all fairly static. We knew that Uzbekistan and Ukraine would be cutoff during the draw (because of their enormous number of entries), but the rest of the countries behaved “normally”.

This “normal” behavior is important to be able to estimate final cutoffs since we could assume that for every 1000 cases, the results would be similar from one year to the next. Based on data available to me at the time, at the beginning of the year I estimated that the cutoff would be in the mid to high thirties for EU. Others were predicting either much higher numbers (like current) or much lower numbers, but the second group (the pessimists) generally were basing their numbers on pessimistic assumptions and misunderstandings, and certainly had no data to back up their claims. However, as the year continued, it became increasingly obvious that were were going to see a low final cutoff. Issued numbers were advancing much faster than should have been possible from a low number of cases processed. So – “people” (meaning the mass of selectees) were behaving differently than previous years.

When the selectee numbers were published three countries (Albania, Russia, and Turkey) stood out as having dramatic increases in their selectee counts.  The data to understand that was only recently published (selectee numbers were only available up to DV2015 until recently, then 2016, 2017 and 2018 entries were published at the same time). With only the selectee numbers to go on, I assumed that some “artificial influence” was at work (as we have seen in previous years where agents registered huge numbers of people as part of a fraudulent scheme). This was referred to in discussions as my “agent” theory. I reasoned that high numbers of cases in those countries would be no shows or non responses because they were fraudulent entries. However, It became clear by Feb/March that the three new “big” countries were being issued visas, so the agent theory was busted.

So what is the explanation?

Well, the answer to that is in the 2016 to 2018 entry data.

I noticed that the “mix” of entries (and thus selectees) changed radically in DV2017 and particularly DV2018. That in turn means the mix of selectees will have changed.

We could see from the selectee count in DV2018 that along with the U2 countries, there were now significantly more selectees in Albania, Russia, and Turkey. However. looking at the data I see also large increases in entries from Armenia, Azerbaijan, Belarus, Georgia, Kosovo, Kyrgyzstan, Moldova, Romania Tajikistan and Turkmenistan. Basically most of Eastern Europe. The increases were sudden and dramatic. All those countries had at least 50% more entries by 2018 as compared to 2016, some of them seeing doubling and trebling of entries in just a two year period. Western EU countries did NOT have similar growth.

Generally speaking those entries seem “non artificial”. My clue to that is if I see a country where the entries are increased rapidly but the derivative rate for the entries drops also, there is probably a scam going on. Only Moldova fits that criteria. DV2018 had twice as many entries from Moldova than it had in DV2017, but their derivative rate went from 1.6 in previous years down to 1.4. It seems someone has accessed a large database of people (probably an educational institution) and registered them proactively. That theory is confirmed by the number of selectees for Moldova. It had 6 times more entries than Romania, but only 3 times more  selectees. The rates of selection should be the same within a region (unless a country is limited like the U2 countries). So – fraud detection methods kept many entries from Moldova out of the lottery winners.

Anyway – other than Moldova, there has been a big shift in the entries, so what does this all mean? Well, there has been a big increase of interest in the lottery from Eastern European countries in the past year or two. That means the “mix” of people within the selectee group has changed. As a group they will behave differently than previous years. They will respond at different rates, show up to interviews at different rates, have different derivative rates, and different success rates. Their motivations for moving from Eastern Europe are going to be stronger compared to selectees from Western Europe. This is not likely to have been a single year change, I suspect DV2019 and beyond will see similar trends (although we won’t know for sure for some time). So – we can expect the DV2019 selectees in EU to behave more like DV2018 selectees than selectees in previous years.

I should also point out that EU was allowed to exceed it’s quota which is a very good thing. That was possible mainly because AS region was under its quota by about 3000 visas (due to the travel ban). If the ban were to be lifted, that could change the DV2019 outcomes.

I don’t have any data to try and guess the final cutoffs for DV2019, so I am not going to answer the inevitable “what about my case number” questions for now.  Please just read the basic questions post and show some patience. It is likely we won’t really know about EU until the final DV2019 VB in July 2019. There are many things that could happen between now and then – so – be patient.