Guild Wars




Further Experiments in Guild Wars GvG

In a previous article I wrote about the current Guild Wars tournament system, discussed single elimination vs. Swiss tournaments and pre-set maps, and gave the math of the Opponent's Match Win Percentage tie-breaker we tested during the Autumn Ladder Season playoffs.

This time I'd like to explain more tournament rules changes we will be testing in the current Winterfest 2006 Ladder Season, such as graduated prize payouts, larger number of entrants, and other changes.

First though, I'd like to answer a question that came in about the previous article.

Xavier wrote in and asked about the 1 point for a draw mentioned in the Swiss scoring system and how that would be possible in GvG, considering we stated that matches that didn't finish by the end of the time limit counted as losses for both teams.

Good question Xavier! It is possible, though unlikely, for a draw to occur in the current GvG system. If both Guild Lords are killed within a very short period of time (seconds), then the game will show a "There was no winning party" message. In a playoff tournament, this would count as a draw and thus 1 match point for both teams. This occurrence is rare but we have seen it happen before.

Also, we use 3 match points as the amount for a victory. Why? If we establish a system where a win grants 3 points, we can consider using 2 points for partial victories or 1 point for a draw in the future, if we need them. For example, a partial victory might be defined as whichever team was winning when a predetermined time limit expired.

After all, it is always better to plan for potential future needs than to constantly change your systems. Now on to the new rules we are testing.

Larger Number of Participating Guilds

One of the advantages of an online tournament is that we don't have to provide a computer for each participant like we would at a live event. And, since players already have their own systems, we don't have to place a time or space limit on the number of competitors.

So why not allow more guilds to compete in playoff tournaments?

As mentioned in the previous article, the current guild ladder system has some weaknesses whereas actual tournament play provides strong build and play tests for the guilds involved. We'd like as many players as possible to have this opportunity. Under our current system we can handle up to 32 guilds, so we want to give it a try and see if there would be that many guilds available to compete at a time.

Graduated Prizes

Almost all of the feedback we received after the Autumn Ladder playoffs was positive. Everyone seemed to like the basis of Swiss play and enjoyed being able to compete against several opponents rather than just one opponent, as in the single elimination system we previously used.

However, a few folks wrote in that if their guild lost the first round or two, they felt there wasn't much incentive to try very hard in later rounds, especially once they knew they wouldn't advance. This is where graduated prizes come in.

If you look at the prize list on the Winterfest Season rules you will see that in addition to breaking out separate sets of prizes for the guilds that make it into the single elimination finals, we have also created separate sets of prizes that participating guilds can win in places 9th through 32nd.

Remember, one of the purposes of Swiss play is to make each and every match a true test of skill by pairing guilds against opponents with similar results. Under this system, even a guild that has lost the previous 4 rounds will still want to play hard in the 5th match in order to try to win more prizes.

Lack of Ladder Advantage in Swiss Play

Another question that came up was about seeding advantage in Swiss play as compared to how it was handled in single elimination.

In single elimination, the best guild plays against the lowest ranked participant, second best plays against second lowest, and so on. This is primarily done to give the higher ranked guilds a better chance of advancing to the final matches. It is a reward for doing well during the season. This is most appropriate when there is a high level of confidence in the validity of the seeding itself.

The current guild ladder system gave a decent approximation of rankings but not a very specific one. Due to the ever-changing nature of guilds' membership rosters, the actual seeding of one guild as compared to another is a little fuzzy outside of a tournament.

So, while we give the same seeding advantage to the participating guilds for the first round of Swiss play, the advantage is gone at the end of that round, and guilds are then paired against each other based upon their match record and tie-breakers from the actual tournament itself.

Remember, the purpose of Swiss play is to pair each guild against opponents of similar strength so that each match is a true test of skill. This is done to provide fun competition for the participants AND to help create a set of final rankings that everyone can be confident in.

Then, when the cut is made to single elimination for the quarter-finals, the eight remaining guilds are paired against each other using the same system of giving the higher seeds a slight advantage by pairing them against the lower-rated remaining competitors.

So, at the end of the five rounds of Swiss play at the close of the Winterfest Season Tournament, the remaining eight guilds will be paired as follows:

  • 1st in standings vs. 8th in standings
  • 2nd in standings vs. 7th in standings
  • 3rd in standings vs. 6th in standings
  • 4th in standings vs. 5th in standings

This gives the first and second seeds (based upon their tournament results NOT their ladder rank) a slight advantage (and the 7th and 8th seeds a slight disadvantage). Note that the 3rd through 6th seeds are still paired against similar strength opponents.


Finally, I wanted to talk a little more about tie-breakers. After the five rounds of Swiss play in the playoffs, it is still possible for a couple of guilds to end up tied, even after factoring in the Opponent's Match-Win Percentage. This is because the system works best when matches are all best 2 out of 3, because each Guild's Battle-Win percentage and Opponent's Battle-Win percentage are used as further tie-breakers.

Since we are unable to run five rounds of best 2 out of 3 Swiss play (it would probably last 13-15 hours), we will be using each guild's final ladder rank as the second tie-breaker, but only if it is needed.

Thus, the rank that each guild finishes in the season ladder will still have some meaning in the actual tournament by affecting the first round pairings AND as the second tie-breaker if needed.

I hope that gives you a good understanding of the rules systems we are testing in the current Winterfest Season.

We are always trying to move Guild Wars to the fore of competitive e-sports and are open to your ideas and suggestions. Please write in to with your ideas.

Michael Gills
ArenaNet Tournament Coordinator