This is one of the biggest problems with a 'raw' points system... especially in the case of MLG. I understand they want to reward players for attending.. but penalizing players/teams that do not have the $$/time to attend more then a single/two events is just absurd. Raw points only works in NASCAR and there it has its flaws. The best system is a relative tournament results system. This system is relative to the results you have in the events you attend. By simply putting some variables in place to ensure repeat attendance (ie: must attend 2 events to qualify for finals - regardless of your finish).. you can ensure a more fair system.

I will post a link (need to upload it again)... to my GPS (Global Points System). This system uses 'Tier' levels to define the 'weight' of events/tournaments and applies the weighted points to each. From there.. a player/team earns points and a GPS Rank% based on their finish against only the other players in the events they attend. This system is a mix of my ( http://www.esportrank.com player rankings) and PGA Golf Rankings. Due to the nature of eSports (with the lack of real league structure).. we more closely represent Golf in terms of competitive structure.

For example..

Idra competes in 5 'A Tier' and 3 "B Tier" events and finishes on average of 6th place in each (with 2 championship wins in the B Tier events)

MC competes in 4 'A Tier' events and finishes with 3 Championships and a 6th place finish.

Using the MLG model and the model of many other eSport organizations.. that use only a 'raw' number system (ie: 1st = 1600 points, 2nd = 1000 points, 3rd = 600, etc...)
Idra would be a much higher ranking player (based on points total).. compared to MC..

However, a relative ranking system would look at only how the players did 'relative' to their specific events. Even if those 2 players never played in the same events. As long at the 'Tier Levels' were defined, there can be a correlation.

Look at it.. MC won 3 of this 4 events.. all A Tier events.. but because he did not 'attend' as many as Idra.. he is ranked lower.

This is a very flawed approach and one that has continued to hinder eSports growth.

Ill get the link to my GPS Ranking system up later tonight.

My 3 cents...

Wren

United States745 Posts

October 17 2011 22:53 GMT

#304

On October 18 2011 07:35 csn_JohnClark wrote:
This is one of the biggest problems with a 'raw' points system... especially in the case of MLG. I understand they want to reward players for attending.. but penalizing players/teams that do not have the $$/time to attend more then a single/two events is just absurd. Raw points only works in NASCAR and there it has its flaws. The best system is a relative tournament results system. This system is relative to the results you have in the events you attend. By simply putting some variables in place to ensure repeat attendance (ie: must attend 2 events to qualify for finals - regardless of your finish).. you can ensure a more fair system.

I will post a link (need to upload it again)... to my GPS (Global Points System). This system uses 'Tier' levels to define the 'weight' of events/tournaments and applies the weighted points to each. From there.. a player/team earns points and a GPS Rank% based on their finish against only the other players in the events they attend. This system is a mix of my ( http://www.esportrank.com player rankings) and PGA Golf Rankings. Due to the nature of eSports (with the lack of real league structure).. we more closely represent Golf in terms of competitive structure.

For example..

Idra competes in 5 'A Tier' and 3 "B Tier" events and finishes on average of 6th place in each (with 2 championship wins in the B Tier events)

MC competes in 4 'A Tier' events and finishes with 3 Championships and a 6th place finish.

Using the MLG model and the model of many other eSport organizations.. that use only a 'raw' number system (ie: 1st = 1600 points, 2nd = 1000 points, 3rd = 600, etc...)
Idra would be a much higher ranking player (based on points total).. compared to MC..

However, a relative ranking system would look at only how the players did 'relative' to their specific events. Even if those 2 players never played in the same events. As long at the 'Tier Levels' were defined, there can be a correlation.

Look at it.. MC won 3 of this 4 events.. all A Tier events.. but because he did not 'attend' as many as Idra.. he is ranked lower.

This is a very flawed approach and one that has continued to hinder eSports growth.

Ill get the link to my GPS Ranking system up later tonight.

My 3 cents...

Cute idea, but (like any power ranking) not useful as seeding for any tournament. Results entirely outside the tournament's scope cannot and should not affect a player's position in that tournament. For example, your system would give HuK a boost for winning Dreamhack Summer, but that cannot and should not be used to seed him at MLG, GSL, HSC, or ANY non-Dreamhack tournament.

ES_JohnClark

United States1121 Posts

October 17 2011 22:59 GMT

#305

That was an example used to rank players not just in 'all' events.. but also JUST MLG Events.. sorry I did not clarify that. It most certainly can be used to rank players in only MLG event. I currently use this very system for my Poker League. With up to 20 players total.. but only around 12-14 on a weekly basis.. this system works perfectly. You must compete in at least 4 of the 10 poker tournaments to have your rank% qualify you or the Finals.

There is no flaw in the system as long as there are certain requirements set in place from the start. (ie. X number of tournaments to qualify, x number of total players required, etc...)

Thanks for the correction.

Wren

United States745 Posts

October 17 2011 23:13 GMT

#306

Very interesting. So results from Dallas (and to a lesser extent, Columbus) would be discounted because of the relatively weaker player pool?

EDIT: looking at it again, it seems fairly subjective again, assuming that event tier is the only criteria. Wouldn't you be better served by a ELO-ish power ranking limited to MLG competition?

Amui

Canada10567 Posts

October 17 2011 23:20 GMT

#307

It's too late to reform atm, but the format does need some changes before next season because atm, the prize pool is too top heavy and seeds are soooo heavily favoured. I understand the need to have seeds play fewer series, but when open bracket players have gone through 5 Bo3's that day, only to run into a top tier player, it really doesn't give much chance.

ES_JohnClark

United States1121 Posts

October 17 2011 23:24 GMT

#308

no.. ELO is not the answer for eSports in terms of 'tournaments'. ELO is best used to rank players in an overall match results system. The www.esportranks.com system that a friend and myself created is a combination of ELO + GPS. The reason that 'Tier levels' are defined for event is as you said.. the level of skill involved. For MLG events however.. its most likely that they would all rank as an A Tier event.. even with the first 2 events being weaker in overall skill. It is possible to 'weigh' them differently/lower if you wish.. and I may prefer that.. but because the format and skill is still very high at all of the MLG events.. it may just be nit-picking a bit too much.

Yes..there is a human element to the GPS and esportRanks.com systems.. but only in terms of the weight of an event. This system can be very restrictive though and should be.. to remove 'bias' from the equation. Defining 'Tier Levels' should also involve more math then just 'guessing or voting'.

motbob

United States12546 Posts

October 18 2011 00:06 GMT

#309

Rain was asked during an interview: If you could tell one person anything, what would it be? And why?

"To MLG Committee: Please change the Providence brackets."

Tektos

Australia1321 Posts

October 18 2011 00:18 GMT

#310

On October 18 2011 09:06 motbob wrote:
Rain was asked during an interview: If you could tell one person anything, what would it be? And why?

"To MLG Committee: Please change the Providence brackets."

HAHAHA

He really did get the short end of the stick, having to fight through Bomber, MC and IdrA if he wants to make it. Beastly T, P and then Z all with good vs.Terran skills.

vVvPastry

United States244 Posts

October 18 2011 00:26 GMT

#311

So glad Incontrol didn't return to afk in pool play...

Vindicare605

United States16095 Posts

October 18 2011 00:52 GMT

#312

I can't wait to see who signs up for the open bracket.

This should be a fantastic tournament, the seeds have an advantage but they're still going to have to play regardless.

sjschmidt93

United States2518 Posts

October 18 2011 00:57 GMT

#313

Silly format. I don't like it.

dcemuser

United States3248 Posts

October 18 2011 01:05 GMT

#314

Well the good news is that because the loser's championship bracket doesn't close up until near the end, the chances of an extended series before the semis are zero.

Korlinni

125 Posts

October 18 2011 01:07 GMT

#315

going to be soo much "pro" talent. I am prob the best SC2 player going to providence.

RedFoxOne

Canada39 Posts

October 18 2011 01:15 GMT

#316

So if I understand things correctly, there will be no pool play (I mean for the top 16 seeds) at Providence? They wait for elimination rounds?

Inside.Out

Canada569 Posts

October 18 2011 01:29 GMT

#317

such a terrible format, the seeding gives waay too much of an advantage

VTPerfect

United States487 Posts

October 18 2011 01:35 GMT

#318

lol.... MLG doesnt reward multiple mlg trips it rewards top 16 at the very first mlg and then riding the ridiculous amount of pool play points all the way to the grand finals. fail

Ownos

United States2147 Posts

October 18 2011 02:09 GMT

#319

LOOOOOOOOOOOOOOOOOOOOOOOOOOL

MC right on Idra's heels. He's probably hoping that Bomber or someone else down the line takes out MC. Haha wow "GL HF my rival" haha