Guild Wars 2

Just do it like hockey.

Keep tiers. Each week, there are 7 checkpoints. Whoever is winning after the first 24 hours wins 2 points, second wins 1 point, third wins 0 points. Repeat for each 24 hour session. After the week ends, tally up the points from each of the 7 checkpoints.

Wipe match points every 24 hours, but keep siege and claims. This means that having one or two great days doesn’t decide the entire week, and provides more opportunities for the other servers to steal points throughout the week. It also prevents the problems with daily resets and enforces the idea that a weekly communal effort is required.

Highest cumulative points moves up, lowest cumulative points moves down, regardless of color or seed number. The goal for each server is to get as close to 14 series points as possible by winning day-to-day. In the event of a tie, coin toss/dice roll (determine winner first, then minus them from the pool to determine loser in the event of 3-way ties).

This takes into account population shifts across the entire week, allowing servers to play slightly better match-ups than if it was winner-up-loser-down. It also allows players to better choose when to play and allows servers to better plan when they need to push hard, which breaks down stress and adds a layer of competitiveness.

Edit:
I recognize that this isn’t exactly the same as how hockey handles things, but a seven-series weekly contest with a somewhat Swiss-style tournament system seems a little bit better overall than random rolls, rating and etc imho.

Might be fun. I wonder if it would make any difference though? Would the same server win every day?

Maybe it would incentivize a server to have a WvW day and queue all maps on that day and try to win a day.

Even if the same server did win every day, it’d break up stagnation by allowing servers to move more often. It’d also provide match-ups with the potential to be much more competitive, as instead of the weekend more or less deciding the vast majority of matches, we’d see servers that have strong late week presence and weak reset night showings getting an opportunity to win overall.

Right now winning on Fri/Sat/Sun usually gives you enough of a points lead that the rest of the week is invalidated. That seems a bit counter to the overall goal to me. Thursday nights are considered throw away nights in 9/10 match-ups due to this. Moving to a nightly reward system makes every day matter in some way.

Edit:
In situations where one server was winning every day, it’d make playing for second more important and a viable strategy. Gaining more points than the third place server would keep you in the tier and give you an opportunity to win again next week.

Did some sloppy math for this week’s Crystal Desert vs Yak’s Bend vs Kaineng. I picked this match specifically because it’s close and will best demonstrate the fluctuations between day-to-day scoring.

http://mos.millenium.org/servers/view/40/

Day 1
2013/06/08 19:54 -- CD: 28,620 // YB: 31,830 // KN: 30,020
CD: 0 // YB: 2 // KN: 1

Day 2
2013/06/09 20:00 -- CD: 29,810 // YB: 30,190 // KN: 31,670
CD: 0 // YB: 3 // KN: 3

Day 3
2013/06/10 20:02 -- CD: 29,700 // YB: 29,740 // KN: 31,960
CD: 0 // YB: 4 // KN: 5

Day 4
2013/06/11 19:58 -- CD: 25,160 // YB: 20,720 // KN: 23,700
CD: 2 // YB: 4 // KN: 6

Day 5
2013/06/12 19:54 -- CD: 31,860 // YB: 29,950 // KN: 28,930
CD: 4 // YB: 5 // KN: 6

Day 6
2013/06/13 19:53 -- CD: 29,710 // YB: 31,630 // KN: 27,950
CD: 5 // YB: 7 // KN//: 6

If the match-up were to have ended yesterday, then even though Crystal Desert was the highest overall scoring server at the time, they’d be moving down a tier as they had the lowest series score. Winning big toward the end of the week wouldn’t have been enough to save them if matches had ended yesterday.

Likewise, Yak’s Bend would have won the series and moved up. They placed consistently better each day than either Crystal Desert and Kaineng, and earned the highest series score. Kaineng would remain in the tier, having come in second place.

When the tiers were switched up, color placement would be randomized. Even though Crystal Desert would have moved down, they wouldn’t have been guaranteed to be the green team. Likewise, consistent Kaineng wouldn’t be guaranteed as blue, and Yak’s Bend wouldn’t be guaranteed as red.

--------------

I understand that this system doesn’t correct all of the problems that a simple winner-up-loser-down system creates, but I still feel that this is a good starting point for discussions.

Perhaps using rating to determine seed and resetting the matches every four weeks to match seed placements might work a bit better. For example, let’s say that in our first week we have 1: SoR 2: BG 3: JQ 4: TC 5: DB 6: FA 7: KN 8: MG 9: YB as our first few NA tiers.

In the first match-up, let’s say that BG and JQ double up on SoR, causing them to lose the series. Meanwhile, FA wins the T2 series and DB loses, and YB wins in T3 while MG loses. Let’s say that CD wins in T4 and moves up.

Our next week would then look like BG vs JQ vs FA, SoR vs TC vs YB, and DB vs KN vs CD. A few more weeks of match-ups occur, and we have our new seeding order. Let’s say that the top nine servers by rating now look like: 1: JQ 2: SoR 3: BG 4: DB 5: TC 6: KN 7: FA 8: YB 9: MG.

After our four weeks of series’ match-ups, our seeded matches would then be:
T1: JQ vs SoR vs BG
T2: DB vs TC vs KN
T3: FA vs YB vs MG

We’d be guaranteed a normal/competitive match-up at least once per month, with the opportunity to experience other potentially competitive match-ups. There’d be some bad ones, yes, but as we’d be collecting points for winning day-to-day we’d have a more realistic view of how servers match up.

We’re normally able to determine which server is going to win a match-up by Monday night with the current cumulative/lump-sum system. There’s not a lot of players out there that aren’t in some way demoralized by 30-50k score gaps after the weekend, and playing catch-up isn’t rewarding in that you’re constantly reminded that you’re behind/losing.

This sort of system would help solve those types of issues. Does it need work? Hell yes it does. Discuss.