Since this is a confusing issue for some, I'm going to copy over an analysis I wrote on our team boards:
Tiebreakers are easy. Basically, if players have the same record, you order them based on their first tiebreaker. If they have identical 1st breakers, go to the 2nd breaker, and if that's still a tie, go to the 3rd breaker. Also, just in case this isn't clear, a win is 3 points, a draw is 1 point, and a loss is 0 points. Players are ranked by total points, with tiebreakers, um, breaking ties. :D
So, if 6 players are X-0-2 (that is, two draws and the rest wins) or better, and four players are X-1-1, the two X-1-1s with the higher tiebreakers will make top 8, and the two with lower breakers will not.
For the technically inclined, first tiebreaker is opponent match win %. If your opponents do well, this number goes up. Thus, being paired against someone with a good record in the final round will tend to improve this # (note that getting a bye is in fact the best possible situation for your tiebreakers, not the worst). So, if you've lost to a teammate, it can sometimes be a good idea to keep playing, if you can win your matches and if their tiebreakers will be relevant. You definitely don't want to play and lose a bunch of matches in that situation, though, as that will only sabotage your teammate.
For example, take the breakers going into the final round of the most recent SCG Richmond, including the best of the X-2s, for reasons we'll see later (Richmond was the last SCG to publish round by round standings):
Op-Match PL-Game Op-Game Matches
Rank Name Points Win% Win% Win% P/W/D/B
1 Cottrell, Scott 13 62.0000 80.0000 57.4286 5/4/1/0
2 Probasco, Andy 13 58.6667 80.0000 55.1026 5/4/1/0
3 Obrien, Adam 13 54.6667 81.8182 55.4662 5/4/1/0
4 Tran, Nam Q 12 65.3333 66.6667 61.7662 5/4/0/0
5 Timoney, Justin 12 60.6667 64.2857 57.3377 5/4/0/0
6 Orlove, Jacob 12 52.0000 75.0000 51.4662 5/4/0/0
7 Early, David 12 48.0000 72.7273 52.0606 5/4/0/0
8 Sears, Van 12 44.6667 64.2857 47.4709 5/4/0/0
9 Halstead, Chad 10 60.0000 63.6364 61.3030 5/3/1/0
10 Pinchot, Jesse A 10 49.3333 53.8462 53.0101 5/3/1/0
11 Michaels, Anthony 10 42.6667 57.5758 48.2222 5/3/1/0
12 Magyar, Peter 9 60.0000 53.8462 58.7273 5/3/0/0
13 Marchand, Chris 9 58.6667 63.6364 55.6177 5/3/0/0
14 Houdlette, Stephen 9 58.6667 60.0000 53.2121 5/3/0/0
Before we start, note that there are only 7 people with breakers over 50. Now, the low breakers will rise a bit because everyone here is getting paired against someone with a good record, but there's basically no way for the 7 people with good breakers to be out of contention if they end this round with at least 13 points. Second and third tiebreakers are essentially irrelevant here (as usual) because we don't have people with identical 1st breakers. Oddly enough, 5th and 6th place after this round will have identical 1st breakers, but neither of them is at risk of not making T8.
Also, I'm incorporating some knowledge of what the pairings were, since that's something you'll know going into the last round.
First, let's look at the lucky players who are X-0-1, or 13 points. Two of them will be paired against one another and will draw in. The third will be paired down. This tournament was unusual in that there were a large number of people with winning records and poor breakers. In fact, since only 6 in-contention people have breakers comparable or better than Adam Obrien's, it doesn't matter which 13s are paired--the remaining one will make T8 win, lose, or draw. Two 13s will draw to 14, at most three 12s will win to 15, and there is one 10 who can win to 13 and have better breakers. That still leaves two slots for more 13s. If the 12s draw to 13, then it's just an issue of breakers, and all the people currently at 13 are still fine.
What actually happened: Brassy was the one paired down, and he chose to play (to try and knock a bad matchup out of T8 contention). The others drew into T8.
Next, the 12s. Nam or Justin will be paired up against one of the 13s (most likely Nam, unless he's played that person already), and can draw in because their breakers are so good. As I said before, there is no way for someone at 13 points and with a breaker above 50 to be knocked out.
What actually happened: Nam played Andy, but didn't realize he could draw in, so he was fine with Andy's decision to play it out.
Next, Justin or Nam (whoever wasn't paired up) will likely be paired with Jacob, unless that match has already happened in the swiss. Assuming it hasn't, they will both be able to draw in, because everyone else has awful tiebreakers. Jacob's 52% will go up a bit because his opponent is 4-1, and the closest behind him is 49%, who would have to have all his previous opponents play in the final round and win for his breakers to rise significantly. Even if they do, that'd put Jacob in 8th place. There's no way for him to not make T8 if he draws. Since I figured this out, I offered the draw and it was accepted.
Now, we know that the two 13s make it. We know that me and Justin make it, and that Andy does. If Nam wins, he is also in. That leaves two or three slots left. David and Van could draw here, but we know that (at least) one of the 10s will win, to be at 13 points. They're at a high risk of drawing themselves into 9th, as Nam winning or two 10s winning leaves them with one slot. Since Van's tiebreakers are abysmal, he can't draw here. David is also at risk if it comes down to tiebreakers, so he too has to play. The winner eats another T8 slot for sure. That means if Nam wins, there will be one slot left.
Now, the low end of the bracket--the 10s. All these players must play out their matches, as they will need 13 points to make top 8. A draw puts them at 11 points, behind the X-2s at 12, and clearly out of contention. Now, Chad and Jesse will probably play here, leaving Anthony paired down. Whoever of Chad and Jesse wins will have better breakers than Anthony, so if he wins he's still only in T8 if Nam loses. The winner of the Chad/Jesse fight should make it in.
Here's where it gets bizarre, though: if Brassy wins his fight and Anthony loses, that's only 7 slots filled, which would let someone at X-2 miraculously squeak into the top 8. Unfortunately for the other X-2s, if Nam loses, he will have the best breakers of any X-2 (especially with a boost from playing against Andy).
So, the T8 is set at Scott, Andy, Adam, Me, Justin, Nam if he wins or if he and Anthony both lose, Anthony if he wins and Nam loses, David or Van (whoever wins the match), and Chad or Jesse (whoever wins that match), all assuming that no one is bizarrely paired due to prior matchups (which they weren't).
Let's take a look at the top 9 players after Round 8:
Op-Match PL-Game Op-Game Matches
Rank Name Points Win% Win% Win% P/W/D/B
1 Tran, Nam Q 15 63.1481 66.6667 63.1502 6/5/0/0
2 Early, David 15 49.2593 71.4286 50.7889 6/5/0/0
3 Cottrell, Scott 14 63.8889 80.0000 60.5915 6/4/2/0
4 Obrien, Adam 14 59.2593 81.8182 61.0783 6/4/2/0
5 Probasco, Andy 13 61.1111 69.2308 57.4306 6/4/1/0
6 Timoney, Justin 13 61.1111 64.2857 58.9423 6/4/1/0
7 Orlove, Jacob 13 52.9630 75.0000 51.0431 6/4/1/0
8 Pinchot, Jesse A 13 49.0741 56.2500 52.7898 6/4/1/0
9 Marchand, Chris 12 58.3333 64.2857 55.6102 6/4/0/0
So Nam won and Anthony lost, leaving exactly 8 players with 13+ points. Because tiebreakers matter, people were forced to play in a way that made them irrelevant for determining who made T8.
Any questions?
