Guild Wars 2

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Week 76 aka post s3 week5 posted

GL

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Ken, at what % of the stat probabilities should a server start expecting to move up or down a tier?

“tiers” really aren’t a thing anymore for matchups. the leader board still has a “tier” column, but really it’s just a “matchup number”. tiers still make sense for ranking purposes, but there’s no probability involved there — after every match your new rating is 100% accurate and absolute in determining your new rank.

so if you are wondering whether you will be moving up a tier or not, the thing to look at is your own server’s rating, and the ratings of the servers in the next tier. if it looks like your will end up with a higher rating than one (or more) of the servers in the next tier, then you can expect to move up.

for matchups, “moving up” or “moving down” really means, “can you expect to play against harder (or easier) opponents.”. for that purpose, I suppose the magic number is 50%. if the higher matchups are more than 50% likely then you can “expect” them (i.e. it is more likely that they will occur, rather than not occurring).

but on the whole, the question doesn’t make a lot of sense. it’s kind of like asking whether you can “expect” to get a 7 when rolling a pair of 6-sided dice. you can expect a 7 all you like but if you roll a 2 then that’s what you got.

-ken

Can someone explain how DB has a 2.5% chance at FA who is only 65 rating points away from them and SBI is at 2.9% when we’re 250 away and DR is 1.9% at 256 away?

Heck YB is less than 250 points from top of the leader board and they show as 0% chance of hitting tier one.

Dragonbrand
17.556897 Yak’s Bend
15.021155 Sea of Sorrows
2.535742 Fort Aspenwood
75.075381 Isle of Janthir
42.996094 Northern Shiverpeaks
42.318688 Crystal Desert
2.917514 Stormbluff Isle
1.578529 Devona’s Rest

It’s more likely to DB to swap with FA than fighting them, that’s why then chance of DB fighting FA is so low. You can see that the chance of DB getting a T2 match is actually 17% from which 15% is YB/SoS and 2% is YB/FA.

About the chance of SBI and DR facing DB: it it because IoJ/CD/NSP can roll lower than SBI/DR (if these two roll high), and that will push these servers on DB matchup. However for YB to get in T1, they need to be within 200 of rating of the 3rd to even have a chance of rolling higher than 3rd (with the 3rd rolling lower than YB).

You should look on the second table instead of the first… I think it’s easier to understand.

Updated scores

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Updated week 7 scores. Happy thanks giving everyone

Fuzzion, i’ve known about these graphs you published but never examined them closely, they’re fascinating, thank you for publishing them.

XacTactX.6709:

Fuzzion, i’ve known about these graphs you published but never examined them closely, they’re fascinating, thank you for publishing them.

Thanks

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Updated to week 79

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Updated to week 80. post-s3wk9.Changes occurring to WvW. I will see how this affects the charts and modify accordingly.

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Updated and have a merry Christmas

Once again ken, thanks for your awesome match up probabilities and merry christmas to you.

Happy 2015 everyone

Added glicko chart and data points.

Have fun and enjoy

sorry about missing last week.

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

May 2015 be an awesome year

Score updated

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Sorry for the delay. Here is the update

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

-Added Score points not from tick with PPK% average

- Week 15 post s3 scores.

- Some minor mistakes were corrected

- Added Fan art by Vorddak

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Snowreap.5174:

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Thanks Ken

As per tradition, Special 16 weeks matchup stats to be presented tomorrow.

Scores may or may not be deleted after HoT expansion if there are drastic changes to the way scores are calculated in glicko.

Ok done. Scores updated and shows last 4 months standings. Nothing frrom Anet if they
are changing glicko, so leaving as is. I wonder when s4 will be announced?
Data taken from both sites – mos and gw2 score

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

scores updated. Some minor mistakes s im too tired to fix :P. Willl do later here is the beta for week 17. I also noticed some sites are outdated. Please give me the new sites or they will be deleted

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

added week 18

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Hi thanks for this, its very cool. Can I ask what the line chart of Score Not from Tick means? Can you explain it?

Is it all non PPT scoring such as Dollys? Or are they still tabulating PPK despite the sneak attack having ended?

For example the last one, WK 89 PPK for SBI is 60 – is that 60%? Because I find it hard to believe that 60% of score was from PPK, even if Dolly’s are included.

primary points from stomps i guess

btw: nice job snowreap, thx for doing this

Before i add the current score I have emailed the website that I took the info from pertaining to PPK. If i dont hear from him then i will remove PPK altogether.

Unless someone knows about the gw2score ppk maths and how its derived?

Fuzzion.2504:

Before i add the current score I have emailed the website that I took the info from pertaining to PPK. If i dont hear from him then i will remove PPK altogether.

Unless someone knows about the gw2score ppk maths and how its derived?

I believe that calculation on gw2score has been broken for some time. As you can see they present some servers with a negative percentage (which is impossible since there is no mechanic in game which subtracts from the point tick).

http://www.gw2score.com/currentscore/total_score/desc?match=1

by request, I’m posting estimated probabilities for the upcoming matchups in this thread also. these numbers are based on the currently predicted ratings (which are in turn based on the current scores), but if the actual end-of-match ratings differ significantly, that will throw the numbers off.

the “na1” and “eu1” files have probabilities of getting a particular server as an opponent, while the “na2” and “eu2” files have probabilities of getting specific pairs of servers as opponents.

-ken

Snowreap.5174:

As always, thank you for the data.

I wonder if Anet cringes every time these numbers are posted?

I doubt it. the system is probably working as well as it can given the constraints it has to work within. “as well as it can” isn’t great, but I don’t see much that can be done to improve things any further without big changes.

I see a lot of suggestions posted, and some of them appear at first glance to be good ideas but when you look at them in detail they either create problems or require big changes. variable map caps, for example, sounds easy but has a big effect on the rating system (changing what ratings actually measure) and of course has the obvious impact of putting more players into queues who were previously able to play.

-ken

As a numbers geek (that keeps track of his servers W/L record against every other server) I appreciate the work you put into this. I happened to notice a couple things on the Post S3 medal count. As of week 89 AR and YB are one short while DB and SoR have an extra.

- In week 86 it appears SoR got credited with both Gold and Silver while AR got nothing. According to the mos site AR beat SoR and DR that week. AR gets one of SoR’s golds.

- Also in 86 DB got Gold and Silver while SoS got nothing and then in week 87 SoS got Silver and Bronze while YB got nothing. In 86 SoS beat NSP and IoJ while DB finished second to FA and ahead of YB. In 87 YB came in behind FA and SoS. SoS gets a gold from DB, YB gets a Bronze from SoS.

Anyway, keep up the good work. It’s pretty neat to see the trends and think about this stuff.

I try my best to keep these accurate (manual input) but i have inbuilt excel macros that will warn me if numbers dont add up. Send me PMs for that week asap if theres a problem. Sometimes i fill these in when Im extremely hungover and numbers are spinning :P

PS. Removed PPK as it was not updating right on gw2score as some of you mentioned.

by request, I’m posting estimated probababilities for the upcoming matchups in this thread also. “na1” and “eu1” files show the probabilities of getting any single opponent, “na2” and “eu2” files show the probabilities of getting a particular pair of opponents.

-ken

T3 looks set for some excitement again.

WOW, Mag is on fire.DH in T3 and T1 getting exciting this week.
Good luck everyone and may Anet release S4 soon.

by request, I’m posting estimated probababilities for the upcoming matchups in this thread also. “na1” and “eu1” files show the probabilities of getting any single opponent, “na2” and “eu2” files show the probabilities of getting a particular pair of opponents.

-ken

Updated. Congratz to the winners.

by request, I’m posting estimated probababilities for the upcoming matchups in this thread also. “na1” and “eu1” files show the probabilities of getting any single opponent, “na2” and “eu2” files show the probabilities of getting a particular pair of opponents.

-ken

have the predictions been posted? I see that snowreap posted within the last couple hours, but I cannot see the post.