Win rate of first team to reach 250 is 86%

Hey pvpers, I've been maintaining this spreadsheet that tracks the win rates of tournaments. Here is the old spreadsheet pre-HoT: https://docs.google.com/spreadsheets/d/1iXf7RxzjOPi3kEnd2_IfCDsUNPh8AlAFgNpgbq4uR0s/edit#gid=1292434431

After a hiatus and returning to PoF, I am maintaining a new one. However, the new data is for HoT only: https://docs.google.com/spreadsheets/d/1cJUHQEIGtsmw9dwW8Eg0yx7YhyXdkcBY7fjkxKDY6BU/edit?usp=sharing

With the new data, I've noticed that the win rate for the first 250 had decreased from 89% to 82%. I believe this reduction was caused by the meta changing, which is bunkers being driven out.

I am curious as to how PoF will change the win rates. I am thinking of adding team comps to track win rates, but I will need more data for PoF. Currently, there is no tournaments that I can watch in the Spectator's POV.

Let me know what other stats you'd like to see. Currently, I have first kill (which has a win rate of 58%) and first two cap (same rate at 58%).

Hey all, I updated the Post-HoT spreadsheet, adding 33 games, making the total 83.

LINK: https://docs.google.com/spreadsheets/d/1cJUHQEIGtsmw9dwW8Eg0yx7YhyXdkcBY7fjkxKDY6BU/edit?usp=sharing

I also added a new statistic that calculates the win chance of a team falling behind the first team with 250 points:

58.82% Chance of winning with 0 to 50 Point Spread Deficit at 25013.33% Chance of winning with 51 to 100 Point Spread Deficit at 2503.85% Chance of winning with 101 to 150 Point Spread Deficit at 2500.00% Chance of winning with 151 to 200 Point Spread Deficit at 2500.00% Chance of winning with 201 to 250 Point Spread Deficit at 250

I will working on adding maps on the data set. I want to determine which maps has the highest chance of a reversal. So far, Foefire and Temple are looking like the prime candidates because of its mechanics (Lord Kill and Meditations).

In an online, tick-based RPG, most "chances" for something to occur are typically assumed as a given if the chance is at or above 75%. It's the reason why things like GW1 Escape melee Rangers and E/D with Mirage Cloak were marked as targets which needed a counter other than just regular attacks (because regular attacks were simply assumed to never hit due to 75% or more chance to block attacks). Ever stack crit chance to 75% or higher without hitting 100%? Still feels like 100% in most circumstances (or it will be strange, to say the least, to see a long streak of non-critical strikes on the screen).

If nothing else, this experiment just proves that GW2 PvP matches last precisely twice as long as they should. Shortening the arbitrary "500 score" win condition to a total of 250 would at least make grinding through PvP matches a lot less of a slog. Aside from just re-working or adjusting individual aspects of the cancer conquest format itself, another way to instantly make the current format more tolerable would be:

• All "ranked" matches now have a win condition point total of 250. Match timer adjusted from 15:00 minutes to 8:00 minutes.
• All Tourney matches will now be determined by a "Best of 3" using the 250 point win condition format. There will be 2-minute breaks between each 250-point match during which players on either team can change builds or swap characters. Teams may also ready up during this break period.
• Baseline respawn timer lowered from 20s to 15s.
• For each point that your team does not own, defeated players on your team respawn 3s faster.

There, now you have a compacted format with more ways to swing games due to respawn advantages (which is something that other games do in King of the Hill or Conquest modes; the homogenized respawn timer is probably the single biggest barrier to comebacks in all of GW2 PvP).

On the post by @Swagg.9236

Win conditions:First to 500 total or a 250 lead.Win at 8 min if there is a 100~249 lead.But if at 8 min none of the above apply then extra time is applied, up to 15 min total game time or a 100~150+ lead or 500 is reached.If at 15 min both teams are equal then sudden death rules could apply. First to get +1 point.

Would the above suffice to shorten blowouts but still allow close games to play out?

Obviously just a rough draft but hopefully still legible enough to decipher the intent.

I think having a reversal mechanic make games more interesting. I love Temple because of the meditation buffs, and it gives losing teams a fighting chance if they are behind. Same with Forest and Foefire because of NPC point kills.

I think they need to design more maps with a point mechanic. Capricorn is the latest, but I have yet to spectate a match on that map to determine its reversal chances.

If anything, I hope the data I presented would help players make decisions in game. In a map where there is a strong reversal chances, the losing team should take advantage of it quickly. Because if they are behind more than 100 points at 250, then their chances of winning becomes very slim.

Hey all, spreadsheet is updated once more. I added about 50 games, bumping the total games at 132.

Here's the link: https://docs.google.com/spreadsheets/d/1cJUHQEIGtsmw9dwW8Eg0yx7YhyXdkcBY7fjkxKDY6BU/edit?usp=sharing

I added a new statistic that shows the reversal chances on each map. There are only four maps in the data: Temple, Foefire, Forest and Kyhlo.

I assumed that Temple would have a high reversal rate, but I found out that Forest sits at the top. Here are the stats:

Forest 22.22%Temple 19.35%Foefire 17.65%Kyhlo 6.45%

I believe Forest ranks the top because of the size of the caps and the secondary mechanic. If you look at Forest, you'll notice that the caps are equal in size. Compare this to Foefire where the middle cap is large. Because of AOE pressures, the caps are easily exchanged in Forest. Finally, Forest has the NPC point mechanic. Since the NPCs are out in the open, losing teams are pressured to take advantage of these NPCs. A timed kill steal would put them in the lead or lessen the point spread.

On Temple, it's surprising to learn that the reversal chances are less than Forest. One would assume that this would be at the top because of the meditation mechanics. But the meditation mechanics itself can be used by the winning team and shut down the losing team's comeback. But Temple is fascinating because of tranquility and stillness. Even if a winning team is in a lead, too much commitment to meditations could lead to disasters such as losing their caps and thinning their players.

Foefire is an interesting map. The Lord is a secondary objective, granting teams 150 if they kill the opposing team's lord. This could turn the tide of matches, but it's also risky since your team will be thinned out from caps. But the main thing that sets this map apart from the rest is its huge middle cap. There is no AOE pressure on this, so the middle can become contested for a while or until the end of the match. Because of this, two caps are crucial. If your team took a two-cap first, you have a 76% chance of winning.

Finally, Kyhlo is the map with an abysmal rate for a reversal. If your opponent is first to reach 250 in this map, you have a 6.45% chance of winning. And it's no mystery because there's no point mechanic in here (point buffs or NPCs). This is the map that determines raw team strength.

Hope this helps. I will be continuing updating this spreadsheet, and I plan on including PoF matches in here.

Love the thought behind this - obviously, the data only supports tourney teams and shouldn't be applied to sPVP ranked leagues or (obviously) pugs. Team makeup will be interesting for sure, although with multiple builds used will have some confounders. Other off-the-top-of-my-head thoughts are: does win rate change if a team achieves a 3-cap, even briefly? Does total amount of deaths of team (not kills but deaths) correlate with losing/winning/point spread? I would think both are common sense "yes" but anecdotally I've seen a couple of teams take 'strategic deaths' and seemingly the death timeout didn't affect their win, and seen teams 3-cap and lose (not just on Foefire)

The first 250 can be used for rank or ATs. If a team is lagging behind at or before 250, they can change strategies and hopefully do a reversal. Knowing the spread between teams at 250 would at least give you an idea whether to be comfortable with a win or worried for a reversal.

It also gives players an idea of what their chances are at a specific map.

I feel three cap is an outlier, so I didn't bother tracking it. Of the games I remember, TCG did a three cap on a lopsided game where the opponent had zero points.

It would be interesting to know the win rate for kills, but tracking that would take a lot of time. If anything, I hope ANet would add points breakdown, which shows how many points were attributed to kills, caps and NPCs.

22.22% is still an abysmally low chance of reversing a match. I still say that matches should just end at 250. This game wastes people's time trying to pretend that it deserves to drag on to 500.

@Swagg.9236 said:22.22% is still an abysmally low chance of reversing a match. I still say that matches should just end at 250. This game wastes people's time trying to pretend that it deserves to drag on to 500.

Another way of looking at it is the more permutations the MM can get through the quicker it settles which is a good thing and in favour of the 250 score limit.

On the other hand the game would be seen as over after the first fight at mid and it'd likely become more of a pve gear treadmill. Red wins w'd be the quickest route.

Interesting, good post. Im going to also start recording this data and see if my results differ in anyway.

Hi all, another update. The win rate for the first team to reach 250 is now at 86%. https://docs.google.com/spreadsheets/d/1cJUHQEIGtsmw9dwW8Eg0yx7YhyXdkcBY7fjkxKDY6BU/edit?usp=sharing

I also added the recent UGO Tournament hosted by @"jebro.6370". There are a total of 11 games for that tournament. 10 were won by the first team to reach 250. 1 was a loss, with a point spread of 2 at 250. This reversal was on Game 1 between "Crystal Desert" and "Mad Kings", the match starts at 2:48:24.

And with new matches added, here is the new stats for the deficits:

56.67% Chance of winning with 0 to 50 Point Spread Deficit at 250
15.63% Chance of winning with 51 to 100 Point Spread Deficit at 250
2.56% Chance of winning with 101 to 150 Point Spread Deficit at 250
0.00% Chance of winning with 151 to 200 Point Spread Deficit at 250
0.00% Chance of winning with 201 to 250 Point Spread Deficit at 250

Doesn't surprise me conquest almsot always snowballs.

Yeah. The more I add games in, the more the win rate increase to the point it's normalizing with the previous spreadsheet I had maintained.

I feel that ANet needs to make more maps like Temple. The thing I like about Temple is that the reversal mechanic is exciting. It also speeds up the game if a team can shut down the losing team with the meditation buffs. But it could also give the losing team a chance if they have the meditation buffs.

I really want to see Capricorn added in the tournament rotation. I'd like to analyze it and see how it stacks up with the other maps in my data.

Pretty nice numbers, though I have to comment on your language.

Your results should read something akin to "x.yz% of sampled teams won if deficit was within D."Just to be precise.

There is always a chance of winning until game is completely over.

Thanks! I will update the language after work.

@Swagg.9236 said:22.22% is still an abysmally low chance of reversing a match. I still say that matches should just end at 250. This game wastes people's time trying to pretend that it deserves to drag on to 500.

These numbers are going to be skewed. Half the matches in a single elimination tourney are in the first round. If seeding is done most of these matches are going to be one sided. If you do it by following a team thought to the finals most of their games they are going to be heavily favored. They probably aren't going to have competitive matches until the semifinals or quarterfinals.

If you want to make meaningful decisions about this you need two teams that are equally skilled.

I approximated what an equally skilled team might be, and that is the point spread of 250 or less at the end of the match. If I factor that in, the win rate would be 73% for the team that reached the first 250.