Drawing the new draw?

Yesterday, we saw the draw for the 2015 Eurovision Song Contest semifinals. After a few countries had been drawn, one thing became clear: whereas the predictability in semifinal two is nowhere to be found, semifinal one is full of expected qualifiers (the likes of Greece, Russia, Romania). That brought me to this editorial: What can we do to prevent predictability and truly break up notable voting allies in the future?
I spent a while thinking on how we can do that. Clearly the pots don’t do what they’re supposed to do. Eurovision.tv stated the following about the dividing of the pots:

To create more excitement, like past years, the 33 countries have been allocated to 5 pots based on voting patterns from the previous ten years.

Let’s have a look at an example pot from yesterday. Pot 4 was the most interesting for me, so let’s have a look. Pot 4 consisted of:

• Belgium
• The Netherlands
• Cyprus
• Greece
• Ireland
• San Marino

So we see two notable ‘pairs’ in this group: Greece and Cyprus versus Belgium and The Netherlands. Yesterday’s draw could have provided the following split: Belgium – The Netherlands – Ireland in semifinal one, Greece – Cyprus – San Marino in semifinal two. Had that happened, there would have been no use for this pot: the notable voting pairs would still be paired together in the same semifinal. Did this pot really create more excitement in such a case? No. Did this pot make voting less predictable? No.
However, the current system works for bigger groups. There is no chance of all Scandinavian friends (Sweden, Norway, Denmark and Iceland) to be drawn together when you have a pot of six countries. You would, as we saw yesterday, get at least one country away from the rest. But that, in my opinion, does not solve the problem. That would still have three of them together, giving them the opportunity to work with their little bloc. The current system does not work against smaller ties. Greece and Cyprus can, in any case in the current system, end up together, which is what EBU do not want. So pots this big is, for me, the biggest issue with the current system.
Now, I also realise that you can’t go around saying a particular system doesn’t work if you cannot provide an alternative. Well then, you could suggest having smaller pots, but you would never truly solve the issue, in my opinion. A smalle pot of, let’s say, four would still have both Greece and Cyprus in it, still giving them the chance to be drawn in the same semifinal. I would call for the EBU to go a step further than pots. Pots do not solve the issue. The step I’m calling for is drawing in pairs.
And what would that mean? Certain voting allies would be paired together for the draw and build a ‘pot’. So for example, Pot 1 could be Greece – Cyprus. One of them would then go into semifinal one and the other in semifinal two. This would bring the lowest risk of predictability and break up known voting allies for sure.

For this editorial, I’ve gone back into voting history since 2008, semifinals and finals included, where I then took the average amount of given and received points per country. A short example: Belgium has had the opportunity to vote for The Netherlands six times (the 2008, 2011, 2013 and 2014 semifinal, the 2013 and 2014 final) and have given 58 points to them in those six opportunities. That averages 9,67 points. The other way around, they’ve received 47 points over six occasions (2008, 2011, 2013 and 2014 semifinal, 2010 and 2013 final), making their received average 7,83 points from The Netherlands. That makes them one of the clearest pairs of the Eurovision Song Contest.
I’ve tried to draw up pairs for yesterday’s draw. That wasn’t always easy, because you can’t have everyone’s ‘first choice’. But every single pair is known to vote for each other. Except for one, which I’ll come to later. First, let’s look at which pairs we’d get, had ‘my’ system been used yesterday.

• The Netherlands – Belgium
• Greece – Cyprus
• Armenia – Georgia
• Romania – Moldova
• Serbia – Montenegro
• Slovenia – FYR Macedonia
• Denmark – Iceland
• Norway – Sweden
• Portugal – Switzerland
• Finland – Estonia
• Belarus – Lithuania
• Russia – Israel
• Albania – San Marino
• Ireland – Latvia
• Azerbaijan – Malta
• Hungary – Poland – Czech Republic

These pairs may look odd in some cases, so we need to look at how some pairs got together. Some countries don’t really have a good old friend that votes for them. Prime example here is Israel, who actually get a fair share of points from France and The Netherlands, but there’s your problem: France is not a semifinalist and The Netherlands has much stronger ties with Belgium. So for Israel, you need to look at other factors: where did they give their points to.

The answer: Russia, Armenia and Romania. They eventually paired up with Russia, which may seem odd as you’d expect a former Soviet state to be close up with Russia…?
That also has a history of its own. Belarus gives most to Russia, but they don’t really get the love back on all occasions, which made the ties between Lithuania and Belarus stronger than Russia and Belarus. And that story basically goes for most ex-Soviet countries – they all have a stronger ally than Russia. So the best solution for Russia is to pair them up with someone who doesn’t have a ‘giving ally’, but still gives them a lot of points: Israel.
A second interesting point is that none of the Baltic states have been paired up. No explanation is needed for Estonia and Finland, who have stronger ties with each other than with anyone else in the region. But Lithuania and Latvia? Wouldn’t it be logical to get them paired up? Yes, but no. This once again goes back to a case like Israel’s: Ireland has one close ally: the United Kingdom. Sadly for them, the UK is not a semifinalist so they cannot be paired up with them. After the United Kingdom, it’s surprisingly Latvia who have given most points to Ireland. Possibly the most surprising fact for me, as I had only seen it the other way around. Latvian and Lithuanian immigrants in Ireland seem to make sure those two Baltic states get a good share of points from Ireland.
That just leaves that one odd ‘pot’ we have at the end. The three central European countries Hungary, Poland and Czech Republic.  None of them have true voting allies and were basically… left. But that doesn’t mean there are no ties between these three countries, so it’s good to have them in one pot. Two of them would’ve gone to semifinal one, one of them to semifinal two.

So that’s a lot of rambling done on how and why these pairs were constructed the way they are.
We now need to look at what benefits this system over the one they currently use. When speaking to people about what I’d come up with, I heard the argument that you would always get two Scandinavian countries together in both semis, same goes for the former Yugoslav countries. And yes, that is true, but unless you work with four semifinals, you cannot prevent that from happening. In the current system, you could of course get Denmark in a semifinal on its own (in fact, that happened yesterday) and you could say ‘See, the system works!’. Well no, it does not because when Denmark is alone, that means you have the other three (Iceland, Norway and Sweden) tied up in one semifinal. So despite the fact that Denmark is indeed broken up from its friends, the rest of the friends are oddly put together.
In a case like that, I still think it is better to break up the closest pairs in such a region (Denmark and Iceland, Sweden and Norway) and have two of them in each semifinal.

The last point we need to discuss is what criteria you use for the pairs. I’ve used semifinals and finals since 2008, but you could easily use a different system. No semifinals would be an example? Perhaps count only the last three years? Or maybe the last five years? I used 2008 for a couple of reasons: it was the first year of our current two semifinal system, but it was also a year where most countries were in the contest (like Czech Republic or San Marino). I have also included voting behaviour in the semifinals, because you would need material to base some irregular qualifiers on (Montenegro, San Marino, Belgium and most notably, The Netherlands). Why most notably? The past two years were the only years since 2008 that they’ve qualified and on both attempts they made the top ten, which means that you may not see a clear voting ally when you just count the 2013 and 2014 final.
To close this editorial, I’d like to call the EBU to have a good discussion over the current system they use. Yesterday’s draw has shown that the used system doesn’t do what it’s designed to do. And whereas my system would possibly cause more pots to be shown on a table, I think it is still a better way to fight voting allies and predictability in terms of voting and results.
What do you think about the suggestions given? Do you agree? Do you have other solutions? Let us know!