When Wimbledon made its draw last Friday, two names appeared near one another: John Isner and Nicolas Mahut. What were the odds of that? Roughly speaking, less than 1%, but not much less.

There were accusations that Wimbledon organizers rigged this, even though the draw was conducted in public (which would require some sleight of hand). However, there’s another vague “pairing” that should be more noteworthy because of how frequently it happens.

Roger Federer and Novak Djokovic in the same half of a Slam.

For years, Federer was either seeded 1 or 2 and Djokovic was seeded 3. In the last two Slams, Federer is ranked and seeded 3, and Djokovic is 2. And yet, we can go back 3 years to find that they are in the same half a preposterous number of times.

**2008**

Australian Open–Djokovic upsets Federer in the semifinals

French Open–Djokovic loses to Nadal in the semis

Wimbledon–Djokovic in same half as Federer

US Open–Federer beats Djokovic in semis

**Tally:** Federer and Djokovic in same half 3 of 4 Slams

**2009**

Australian Open–Djokovic and Federer in same half

French Open–Djokovic and Federer in same half

Wimbledon–Djokovic and Federer in same half

US Open–Federer beats Djokovic in semis

**Tally:** Federer and Djokovic in same half 4 of 4 Slams

**2010**

Australian Open–Djokovic and Federer in same half

French Open–Djokovic and Nadal in same half

Wimbledon–Djokovic and Federer in same half

US Open–Djokovic beats Federer in semis

**Tally:** Federer and Djokovic in same half 3 of 4 Slams

**2010**

Australian Open–Djokovic and Federer in same half

French Open–Djokovic and Federer in same half

Wimbledon–Djokovic and Federer in same half

**Tally:** Federer and Djokovic in same half 3 of 3 Slams

## Total

In the last 15 Slams, Djokovic and Federer have been in the same half 13 times. Now what are the odds of that?

## tim

are you guys all retarded???

1 Is always in the same side of the draw as 4..

The same as 2 and 3.

They do this for pretty much every single knock out sporting competition so that you would ‘expect’ that 1 plays 4 in semis and 2 plays 3.. then so 1 plays 2 in the GF

## mat4

It is worse than that: actually, the fedal final was protected by the semi Nadal – Djokovic at Roland Garros. There was one exception: in 2009, when the french organisers opted for a final Nadal Djokovic, after they dominated the clay court season, and having in mind Roger defeat in the final the previous year. Practically, in 14 of the 15 slams the fedal final, or most probably Roger Federer’s place in the final, was protected, and in all those 15 slams, the semi-final draw was fixed. The odd is 1: 4096 for this.

I am sure of if. I manage to predict the semi-final draw for 7 times in a row now, and I can post a link where to find at least one of this prediction. I predicted also that Murray and Sod will land in the same quarter when Soderling was 4th on the atp ranking list.

The differences between the four top players are very small, and small help can change the complete picture. What would have happen if Roger and Rafa played the semi at RG? In 2008, who would have won the USO if the semi were Federer – Murray and Nadal – Djokovic. The result would perhaps be the same, but I would have given Murray much more chances to win in the semi than in the final.

## mat4

I presume that it is not easy to write what seems to be obvious from analyzing the stats: it could perhaps provoke a lot of controversy. I congratulate you to point to the hidden side of our favorite sport.

## Lawrence

It’s a .32% chance.

## Lawrence

If it is truly a 50/50 chance each time, then the odds are .32% so it’s obviously rigged so there is a chance for a nadal/federer final than seeing them meet in the semis.

## Justin

Actually, for the vast majority of these slams, Federer was ranked #1 and Nadal #2, so they cannot be in the same half of the draw. 1 and 2 play either 3 or 4 in the semis. The only potential rigging claim here would be that Djokovic should have been in Nadal’s half more often, and Murray should have been in Federer’s half more often. I don’t buy that someone would bother to rig this, as this isn’t as compelling to rig the bottom half of the top 4.

0.32% is the probability that 13 of 15 times they are on the same side. 0.36% is the probability that they are on the same side 13 or more times.