Risk FFA tournament ranking

[mighty meadow]

I have created (with a lot of help) and planning on maintaining a live ranking of player based only on FFA tournament performance

https://docs.google.com/spreadsheets/d/1GWwhS0I7IRmmfvEw4LedYHFc1Fb_Fao5idrwHfKK6GM/edit?usp=sharing

this is a part of a more vague vision of mine of creating a tournament "Tour" like in the ATP tennis system which I like

https://docs.google.com/document/d/1N6eRjvha16vO0RhUEN520_MjxgC5Ej2h1I-g10XtzfQ/edit?usp=sharing

happy to hear your thoughts on the matter

[hollow spear]

Impressive you’ve managed to compile results from different servers.

Almost all points on the sheet are earnt by winning tourneys, so the hundreds of names of those who got 0 points or almost 0 points, aren’t of much interest, for the amount of time they must take to put together

As a result it looks like only the most recent couple of seasons have been counted, despite there being halls of fame for far more ffa and gm seasons that aren’t covered.

I’ll have a go at making a copy that saves time by only needing names of finalists/winners, so can more efficiently cover the full dataset, rather than just this season

[eternal light]

@hollow spear It's supposed to include only tourneys during the last 365 days and it updates live

That's the reason for it missing many tourneys

[hollow spear]

Yeah, I’m not saying it’s a mistake, just a difference in perspective

Even with those limitations, it’s still probably the most robust measure we have available of FFA skill, arguably even better than FFA rank, because tourneys are played under controlled conditions

[hollow spear]

Official FFA + GM Tourney Hall of Fame
Calculated by pasting into ChatGPT

Pete (5 times)
Kylted (3 times)
Tim (3 times)
UnderTheGun (3 times)
Maren (2 times)
Sterling (2 times)
Parlox (2 times)
Great Khan (2 times)
Zombought2knees (2 times)
Amerigo (1 time)
AceroMike (1 time)
Frail Waif (1 time)
Fionn (1 time)
Velocity (1 time)
Curlza (1 time)
OliveXC (1 time)
etc.

Including Legends Winners
Pete (5 times)
UnderTheGun (5 times)
Kylted (4 times)
Maren (3 times)
Tim (3 times)
CosmicCowboyYT (2 times)
eG Sneaks (2 times)
𝕋𝕣𝕦𝕥𝕙 ℙ𝕠𝕨𝕖𝕣𝕤 (2 times)
Oneshot (2 times)
Sterling (2 times)
Parlox (2 times)
Great Khan (2 times)
Zombought2knees (2 times)
Etc.

Looks like @foggy verge and @versed coyote are the best FFA players according to AI

[mighty meadow]

LOL i am AI expert so don't count on those too much

[hollow spear]

It's just counting how many times their name appears in the halls of fame

[mighty meadow]

@hollow spear i want to give every cutoff made some points and semifinals do add up. and last 365 days make it fresh and more exciting

[mighty meadow]

o yea it will work there

[hollow spear]

Yeah, there are definitely benefits to the way you've done it, and there are benefits of other metrics too. Mainly comes down to individual preference on how you like to see the results presented

[hollow spear]

All time Hall of Fame Appearances FFA (Official +GM +Legends +Zombies +Assassins)

6x
Pete

5x
UnderTheGun

4x
Kylted

3x
Maren, Tim, Xerokee, Sterling

2x
Ragingdeano, [TBH] real_llamas, Sir Tyler, CosmicCowboyYT, eG Sneaks, Parlox, Great Khan, Zombought2knees, Tyler

1x
INNOCENT KITTEN, Flacko, Baldy Kha, Walker Khong, Crackle069, UnderTheGun, Kylted, Amerigo, AceroMike, Frail Waif, Fionn, Velocity, Curlza, OliveXC, TheBull, Giga4man, Sam, Zebia, CoolTim, jjbruh4, 3irc, Silent Legend, IamMomba, Klonaway, Mounteverest, Qlwl, bigdog, ChampionEver, Resch, That Guy, MapleSpace, MozesIronBear, Folgore_Heroes, General Vadar, xerip, Doctor Moros, Troy Bond 28, Walton Judgement 15, Zak, IHitBack, Merl Musket 28, Da Reel Deel, Moly, Hayden Cook, THE AWESOMEST WINNER, Cherry, Legatus, SynnTheDark, appiejan, KingJohn, fanti13, Pierre Stonefield 10, Amoo Bao, Mstigeress, General Vanman 663, Spud

0x
30 million+ Risk accounts, including inactives

[novel trellis]

I should be 2x

👀

I made FFA Finals and GM Finals

xd

Oh 1-3?

Confused what you mean by hall of fame appearances lol

[hollow spear]

It's simply counting how many times your name is in #🔒│hall-of-fame

[novel trellis]

Well yeah I was confused what you mean by hall of fame appearances since I am not sure how other tournaments do their hall of fames.

I don't have time to try any more in tournaments

otherwise I'd be 5x

xd

Sterling is typing a lot so to clarify this is a joke

[hollow spear]

@south fox based on the table it made me, feeding in the roar lists from all 1v1 halls of fame, ChatGPT has named you the best Risk player

[south fox]

Cool

Don’t think risk of the hill should count as that is so small

5 a side is also a team event

You can individually have a good season but not end up winning

It is individual 1v1 and 2v2 tourneys that I think really exemplify skill

[hollow spear]

Yeah, but need enough results to make it interesting, and doing something more complex with the results would take forever

[south fox]

It is good for a first take at least

To be clear this is outside of FFA which is the most popular mode

[hollow spear]

Yeah, here’s the list for FFA

[south fox]

That is probably me as I placed top 3 in 3 1v1 tourneys and 3 2v2 tourneys. No one else has done this

@hollow spear should I count on the FFA list

My one finals was a hybrid tourney

The fact that I am top 50 FFA officially is also hilarious

According to these metrics

[hollow spear]

To keep it simle and objective I included the entire hall of fame for every FFA event with one available.

If I started picking and choosing what matters and what doesn’t or how much something matters, it just becomes a measure of my own opinion, and subjective

[south fox]

I made one semis last year. I don’t think that qualifies as top 50

But yeah what you say makes sense

[hollow spear]

The reality is every player knows the deep details of their own matches far better than the hall of fame does, so for analytical reflection, individual experience always takes priority.

However these punchy simple stats make for nice headlines, that nuanced memories don’t

[mighty meadow]

@south fox who do you think should be higher then you and why?
my rankings are clear and based on last year performence you should count as top 50.
i would love to hear what do you think is wrong with the methodoligy

[foggy verge]

Olive just think's he's lousy at FFA

[mighty meadow]

@hollow spear i kinda wanted a discussion on what i did no offence but you did made this thered about something else .
i would like reaction and criticism

[foggy verge]

Yeah the false equivalency between FFA and 1v1 dies hard

I like what you did Sabr

Personally I would score tournament finalists highly

whether or not they won or placed 'in the money'

[mighty meadow]

@foggy verge finalist are currently get 166% what semi finalist get. (you think this should be 200% ? more?)
i did add some discrepency in the gm for 2nd and 3rd becausr they won a new game.

[south fox]

i just played in a lot of tourneys

i was bound to do okay in one of them statistically due to luck, but overall I get eliminated early most of the time

if just one semi finals gets you top 50 that just seems high to me

i feel there should be more weight for the official risk FFA tourneys over side tourneys

just my thoughts

[south fox]

i am not lousy. I am GM level. I am just not a top 50 in the world GM level

[foggy verge]

You're too humble

I would score all tournament finalists equally

they fought to make it to the top of a tourney

then someone won a game

[mighty meadow]

the weights of tournements was average respons of 16 top players i asked. 👍

[mighty meadow]

i understand you think winner is too high in comperison to finalist

[foggy verge]

It's a big of a sampling problem right

how do we get a large enough sample of a newer players performance to determine how good they are

[mighty meadow]

i guess we wait a year unfortunatly

[foggy verge]

because of the very high variance in the game scoring players based on semifinals and finals appearances is how I would do it off hand...

[foggy verge]

Well with the proliferation in the tournament scene there will be more data

so that's a good thing

[eternal light]

Agreed, I was very surprised when I saw some people had voted GM to have more weight than World Championships.

[foggy verge]

@eternal light isn't a tournament of all GM's more challenging than an open tournament?

[mighty meadow]

who is the owner of gm discord ? @neon tendon \ @hollow spear ? both?

[foggy verge]

And myself

3 people

[mighty meadow]

how technicly addept are you from your streams i would hasitate

[foggy verge]

I am merely Parlox puppet

If you want data I'm sure Sterling would share with you

He's historically been very interested in the same question you're asking;

How to get a fair measure of a players ability in the game

[mighty meadow]

no, i want to make the discussion there . like in this forum with threads about certine question and maybe voting on them

[foggy verge]

Ah yeah do so

That would be well recieved

I really like what you're doing with this

[eternal light]

In my opinion it's more impressive and should have more weight if one is able to score consistently in an open tournament than if they place high in an only GM tourney. When GM's are put next to each other the luck factor rises immensely. Just my thoughts tho.

[foggy verge]

Smg cannot be relied on to provide us with a meaningful ranking system in a reasonable amount of time we have to do it ourselves.

[foggy verge]

That's funny I see it the other way. In an open tournament any noob can ruin your tournament at any step of the way. Increasing variance. Whereas with gm being the floor the mistakes made in games tend to be of a different nature.

The average gm doesn't play close to optimally

[eternal light]

The luck factor still rises because the players are closer to the same level.

[foggy verge]

You're talking about the luck factor within a given game

But not over the average of the tournament

[eternal light]

It's a skill on it's own to learn to play around noobs, the tourneys just include different kind of skills

[foggy verge]

Put it to you this way. You making the semifinals of a 1000+ player event has got to be significantly luckier than you making the finals of a gm tournament with 100~ players yes?

[eternal light]

I think making Semis in WC is less lucky than making Semis in GM tourney.

I might be wrong though, I'm open to change the way I'm thinking.

[mighty meadow]

@eternal light @foggy verge GM you get 3 chances that remove some of the luck component.
I think GM is more skillfull
but the WC is more prestigous and should account for more

[neon tendon]

Gm tourney effectively has 15 rounds so theres more time to prove your skill

In regard to prestige, thats just an opinion. Id argue when it comes to difficulty the gm tourney is more prestigious, if youre talking just about popularity though, thats not what we're measuring here.

[hollow spear]

I’m impressed by your work’s thoroughness, and think it’s the best compilation sheet of FFA results available, and will continue to be, so long as it’s kept updated

The artificial variables of point scale and time cutoff intrinsically have to be arbitrary, even if there’s some agreement on where those arbitrary lines should lie.

My reaction yesterday, was to counter the results, by removing time cutoff and point weights, to compare the rankings

Reassuringly we still found the same top 3 players: Pete, UnderTheGun and Kylted

[mighty meadow]

@hollow spear i think i understand now .
in my opinion removing time cutoff and point weight is also and arbitrary and is equal to set the weights to same for all tournaments( except chaos i guess which you didnt count) and setting the time cutoff to infinity. i dont think those are "natural" choices that are better then artifical ones.
espically the time cutoff is importent as it differantiate between discussion of current best player to greatest of all time player(GOAT) which is different kind of ranking

[hollow spear]

Yeah, all time and unweighted aren’t perfect either, but it illustraits that your preferences of weights and cutoff aren’t the only possible way to handle it.

If you are including arbitrary time decay and arbitrary weights, it could be worth calculating them based on something a little more fundamental.

• E.g. if we agree that a more recent win is more important than a win further in the past, it would be smoother and more accurate to measure time-decay continuously, with results decaying gradually over days/weeks, until they reach 0, rather than having 100% weight for 365 days then discretely jumping to 0% weight from a sudden absolute decay

• It’s clear where you’re coming from on basing weightings between the events on the subjective opinions of 20 top players, but there’s an irony to creating a list of top players, with a method that depends on already knowing who the 20 top players are. This can give an impression of bias to outsiders

The ideal would be to calculate weightings based on fundamentals, on a set of objective numerical criteria, rather than ad hoc opinions.

Tournament characteristics such as number of participants, previous performance of opponents, game results required to reach the milestones, number of games, could be used as numerical inputs for a weighting function, to make results of different events compatable

The golden standard for working out the weights, would be an objective numerical calculation method based on publicly known characteristics of the events, that would give the same output weights for the same input data, regardless of the opinion of the person calculating it

[mighty meadow]

damn i need a computer to answer this

[hollow spear]

The parameters for time-decay, event weightings, and placement weightings must be arbitrary.

It’s essential these are set at a level that’s well justified and defensible, so results are accurate, reliable and unbiased.

The options I suggested aren’t the only, or even necessarily best way to achieve this, but however you do choose to do it needs a highly convincing, and hard-to-debate reason, that goes beyond setting them based on what sounded cool at the time

[mighty meadow]

ok ill try to answer a bit .
even if i weight based on parameters such as player count there wull be even more cofigurabel parameters thats why its simplier to ask diractly players that have the most expiriance with the tournaments which are the top players.
about the year cutoff it is very standart cutoff in many sports and esports. while decaing over time in a more continues way is better its more complex and less understandable and ignors the timing of different settings of tournaments that some players will be better at one not the other.
on a more general side .
1.this ranking are not ment to measuer skill exactly more like form or performance.
2. i want this ranking to have an impact on players behavior such as sign up for more tournaments and not rest on past results etc.
3. while definatly not exact science it is good enough and simple enough. and if it will be widely adopted and referanced ill be happy even if they are not completly optimized.

i want grow this aspect of risk and i belive this is an essintial part. hope the community will embrace it as "official" in some way. so players will have some braging rights and also a tool to asses themself and their opponents.

[hollow spear]

Standardisation across sports, and efficiency of calculation explains the time decay

How did other sports weight their different tourneys? Was that by asking their top 20 too? And how did you identify the top 20?

Guessing your weighting for 1st, finalist, top 16 etc. is based on how another sport or world tour method did it too?

[mighty meadow]

i havily basing this on ATP ranking of tennis

[hollow spear]

Sounds well established, but I’m not familiar with the details of it.

Do they make weights between tourneys by asking the top 20 tennis players?

And use the same weighting for 1st, 2nd, 16th etc. you used?

[mighty meadow]

i dont know how they established weights i think history has a part in it probebly a central body . but they rarely change.
the 1 and 2nd they have a 0.6 multiple for each round( if you go down ). so i did the same for winner finalists and semi. from there i did a multiple based on proportion of players cutoff.
and i gave 0 for not making any cutoff. i belive tennis give some for apperence

[hollow spear]

It’s a shame tennis’s tourneys are 1v1/2v2.

Would’ve been handy to see how they handle FFA

[mighty meadow]

the closest game i can think of is catan maybe or ffa strategy game like aoe or starcraft. this ranking is realy hard. not only that .the tournament rules themself are even harder to make and viewer expirance aswell. but we can do it

[hollow spear]

Yeah, designing optimal Risk tournaments is hard for a few reasons:

1. Unlike races with cars, simmers, runners, horses, greyhounds etc. you can’t time an objective qualifying lap for Risk, to initially order players. Each Risk win/loss is only ever relative to your opponents.

2. In 1v1 games/sports like Chess, Go, Othello, Boxing, Tennis, MMA, Wrestling etc. if you win it strongly implies you played better than your opponent. In Risk there’s such a high luck element, it’s normal to frequently see the less skilled player get lucky and win

3. In FFA it’s not just your opponent’s skill that can make you lose. It’s even possible for an opponent to make a mistake that makes you lose, by allowing a 3rd player to kill you

As a result it’s only ever possible to estimate a Risk player’s skill over a very large number of games, because factors other than individual skill overshadow skill so hard in the short run

[scenic tree]

This!!

[gilded wyvern]

Which 2 am I in? Because I think I should be in 3, FFA (Q2 2021), assassin's S2, GM S3

[hollow spear]

Not sure, spent like 15 mins making that, 6 months ago, and haven’t looked at it since

[gilded wyvern]

Oh lol I didn't realize this was old, I just looked up some old name and it's only instance in the server was here
But now that I looked into it, I can't seem to find the hall of fame in the GM server, I'm pretty sure there used to be one

[hollow spear]

I wonder what happened to it