I believe the best method of group play is a round robin with best of twos instead of best of threes Wait don't run away yet. It's not that crazy. Let me explain.
I don't need to remind everyone of how many idiotic situations we have found ourselves in when there are tiebreakers in tournaments. There can be mass confusion about the tiebreaker rules or simply a bad taste left in your mouth when a player is out of the tournament because of a convoluted map score calculation.
I will use some examples from Dreamhack this weekend to illustrate my point though there are many more from previous tournaments. I want to be clear that I am not bringing this up as a way to rally behind a player that was eliminated. This thought process could be applied to any group. I am trying to show that there is a much better option than what we are doing, though it sounds unconventional.
This was white-ra's group:
The thought process behind White-ra's elimination is that his score differential was less than the other two. AKA: he had more losses or less wins or both.
I feel like this is a very bad way to do it if the intent of the groups and tournament is to find the best player because in best of threes, some players play less maps.
White-ra played less maps. (7 vs. MC's 8). He is eliminated because, with equal losses, he did not win a fifth map and MC did. However, White-ra did not play the eighth map that MC did. There is no indication that white-ra could not have won the third map in the series he lost if it was played once he was down 0-2. The difference comes from the order the maps were played. If they had played the maps in a different order, In the series MC lost 1-2, he would've lost 0-2 and would not have ever had the opportunity to score a map win on the third map. There is no way to know if this is what happened to white-ra or not.
We are comparing 0-2 to 1-2. We are comparing fractions without a common denominator. We are comparing MC's score over 3 maps against White-ra's score over 2 maps. White-ra was not given the same room to show his skill as MC. This would be less of an ordeal if every sc2 game was played on the same map, but it is not.
The essence of what I am saying is that you cannot compare map score if the players are not playing the same amount of maps. It becomes a meaningless and arbitrary comparison.
If you decide tiebreakers on map score, either every series has to be played to three maps or they need to play best of twos. so that the players have the same amount of maps played at the end.
I think playing the third map after the score is 2-0 has it's problems. The third match would be unexciting for both the spectators and players. Additionally, it provides an opportunity for the winning player to go 3-0 and greatly inflate his score by being better than a single player in the group.
As a result, I think best of twos are the best option. I think this method has been avoided because spectators need the feeling of one player "winning" the series. If we can get past that I think it will improve the integrity and fairness of competitions.
I most likely did a poor job explaining my point and turned some readers minds against the idea from the get-go but I strongly recommend taking the time to consider the possibilities and consequences of having best of twos.
Please don't focus on white-ra and MC when considering the idea. This happened to NightEnd a few IEM's ago. It's happened to a number of players in a number of events. Players lose three way ties based on the the third order of tiebreaker rules. We would not need more than one tiebreaker rule if players played the same amount of maps.
MLG did a bo2 online event this year (or late last year?) - the first thing with the KeSPA pros vs eSF guys. I think the cons (really awkward for spectating, not having a clear winner in many matchups) vastly outweigh the pro of trying to avoid this specific situation. These tie breaks are not bad scenarios assuming they are handled correctly, but on occasion an event will bumble them sometime and make it look like another option is more attractive when the reality is proper procedures should juts be in place for if this happens.
If you want to go this way, wouldn't it be better to play all three maps in a bo3? At least in this system you get to always declare a winner, whereas with Bo2's you dont.
This seems like a precarious situation that you intend to fix by making all the very regular situations more awkward. I understand that this might fix this specific situation, but I don't think it would be best for the norm. In general what Heyoka and Plexa said are right IMO. Though this is an interesting idea.
I understand that in this situation if White-Ra beat SortOf in the last map, then it would be a 3-way tie at 5-3 a piece, but it would be quite dreadful to the viewers to have to sit through third games after it has gone 2-0. My idea would be to play the 3rd game of the series off-air. That way it wouldn't burden the spectators and you get the fairness that you need.
On April 27 2013 10:50 Plexa wrote: If you want to go this way, wouldn't it be better to play all three maps in a bo3? At least in this system you get to always declare a winner, whereas with Bo2's you dont.
I'm inclined to agree with what you're saying. I suppose it becomes a matter of personal preference. Is a game three in a 0-2 series exciting enough to justify that approach over Bo2? If so, I would agree it is the better way to do things. Either way I feel like it's better than what we're currently doing. I'm not sure why it hasn't been discussed seriously before.
On April 27 2013 10:57 Chairman Ray wrote: I understand that in this situation if White-Ra beat SortOf in the last map, then it would be a 3-way tie at 5-3 a piece, but it would be quite dreadful to the viewers to have to sit through third games after it has gone 2-0. My idea would be to play the 3rd game of the series off-air. That way it wouldn't burden the spectators and you get the fairness that you need.
I think this is an excellent idea.
I guess the core of the issue is that using map score as a tiebreaker is flagrantly unfair if the players havent played the same amount of maps. From a pure fairness oriented perspective, I would be more okay with the current setup if tournaments allowed tiebreakers more easily and would not do things like decide a players fate on strange head-to-head map scores in isolation like we've seen before. From a spectator standpoint, more tiebreakers would certainly not be preferable for most but I don't think we should be satisfied with the status quo tournaments have seemed to settle on.
I think unless you have a very large group (e.g. The International 2, though Dota 2 has a more solid reason for using BO2, i.e. spawn sides), BO2 just ends up creating more awkward tiebreaker situations than it resolves.
well i disagree, because in a bo3 format, if you win a series 2-0 lose a series 1-2 then win a series 2-1 you have beaten 2 oppoents and been pushed to a final game and pushed someone to a final game giving u a game score of 5-3. If someone else in the group went 0-2 2-0 2-0 they have 4-2, which is the same score difference but they've actually won less games. You could argue that they are the better play because they won their 2 series 2-0 but they also lost 2-0 and you didn't at any point so you have a legit claim to say you were the better player while the other guy also has a legit claim by saying "i didn't drop any games in the series I won" granted this is a problem.
In your example Whitera won 2-0, 2-1 and lost 2-0. Mc won 2-0, 2-1 and lost 2-1. So the series they won they leveled out on, both going 2-0,2-1 but in the series they lost whitera lost 2-0 while MC lost 2-1. So by losing 2-1 MC performed better than Whitera, even though he lost to Whitera. Whitera lost to Sortof 2-0, MC beat Sortof 2-1, again MC can be easily argued to have performed better. I can find a few different ways to skin the cat, they all come out with MC going through and Whitera going out. There is no situation in which Whitera goes through unless you use Head2Head Score as your first tiebreak, which is a bad way of doing things.
If you have bo2's you will end up with a whole bunch of people with 3-3 scores and no clear winner. You are actually making the problem worse. There is no clear winner in a bo2 unless someone goes 2-0. bo2 works in most sports because within each game their is another scoring system. Like in soccer, you play the same team twice in your group and win or lose 3-1, 2-1, 1-0 whatever. If two teams end up with the same amounts of wins/losses they then use goal difference, in exactly the same way as game win/loss is used in groups in sc2. So if we go to a bo2 format we then need to work out a new scoring system to fill in the gap, how do you score an sc2 game?
You certainly can compare map scores when they aren't playing the same number of maps because you use the series wins/loss as your first measure of rank, then you use map score as a tiebreak. In other sports you use game win/loss to determine rank and then goal/point difference to tiebreak. It is the exact same thing. I fail to see how it is unfair. If you can always push your opponents to a map 3, even if they are better than you, you have performed better than someone who always loses 2-0. It would be extremely unfair to NOT count those maps.
On April 27 2013 10:59 synapse wrote: I think unless you have a very large group (e.g. The International 2, though Dota 2 has a more solid reason for using BO2, i.e. spawn sides), BO2 just ends up creating more awkward tiebreaker situations than it resolves.
The International 2 for Dota 2 is an excellent example of what I am getting at.
If everyone is predicting the core issue with bo2's is a lack of excitement over a clear winner, does that make it the case? I agree it is a very possible problem, but could we not try it out at a major event to see how it goes? I did not find the Bo2 format to detract from my viewing experience at the International at all.
On April 27 2013 10:59 synapse wrote: I think unless you have a very large group (e.g. The International 2, though Dota 2 has a more solid reason for using BO2, i.e. spawn sides), BO2 just ends up creating more awkward tiebreaker situations than it resolves.
The International 2 for Dota 2 is an excellent example of what I am getting at.
If everyone is predicting the core issue with bo2's is a lack of excitement over a clear winner, does that make it the case? I agree it is a very possible problem, but could we not try it out at a major event to see how it goes? I did not find the Bo2 format to detract from my viewing experience at the International at all.
They've tried it, it didn't work. People complained about lack of excitement and that it was an unfair system. They had to try out new ways to tiebreak, like a coinflip. It caused and will always cause even more tiebreaks. Playing an even number of games will always result in more ties unless you get very very lucky. The best solution is to use the GSL format and make groups a small double elim bracket.
In dota they can use bo2 because they have other scores to use as tiebreakers such as:
time score - often in a group they use time score (how fast you win or slow u lose) kill score - pretty obvious what this is
we don't have that in sc2. We can't use units killed as a score, or workers made or whatever. You can't even use time score because different maps and races cause games to be completely different lengths. You can't compare a TvT on Whirlwind to a ZvP on Neo Planet S, for example.
On April 27 2013 11:04 emythrel wrote: In your example Whitera won 2-0, 2-1 and lost 2-0. Mc won 2-0, 2-1 and lost 2-1. So the series they won they leveled out on, both going 2-0,2-1 but in the series they lost whitera lost 2-0 while MC lost 2-1. So by losing 2-1 MC performed better than Whitera, even though he lost to Whitera. Whitera lost to Sortof 2-0, MC beat Sortof 2-1, again MC can be easily argued to have performed better. I can find a few different ways to skin the cat, they all come out with MC going through and Whitera going out. There is no situation in which Whitera goes through unless you use Head2Head Score as your first tiebreak, which is a bad way of doing things.
I think you are missing my point. You are saying white-ra deserved to be eliminated because he lost 0-2 and MC only lost 1-2. My point is that White-ra could have lost 1-2 as well if a third map was played. A third map does not get played under the current system and I believe that is unfair because it is boiling down to the order the maps are played and not the number of maps a player could win (thus not effectively eliminating the worse players in all cases).
Hypothetically: If white-ra is unbeatable on map 3 and cannot win on maps 1 and 2 and MC is unbeatable on map 1 and cannot win on maps 2 and 3 With all other variables and group performance held equal, MC advances 100% of the time and is considered the better player. In a different tournament held the very same day with a different map order, White-ra advances 100% of the time and is considered the better player.
If the goal of the tournament is to crown the best player as the champion, I do not think our current system is effective from a purely 'competitive integrity' perspective.
On April 27 2013 10:59 synapse wrote: I think unless you have a very large group (e.g. The International 2, though Dota 2 has a more solid reason for using BO2, i.e. spawn sides), BO2 just ends up creating more awkward tiebreaker situations than it resolves.
The International 2 for Dota 2 is an excellent example of what I am getting at.
If everyone is predicting the core issue with bo2's is a lack of excitement over a clear winner, does that make it the case? I agree it is a very possible problem, but could we not try it out at a major event to see how it goes? I did not find the Bo2 format to detract from my viewing experience at the International at all.
They've tried it, it didn't work. People complained about lack of excitement and that it was an unfair system. They had to try out new ways to tiebreak, like a coinflip. It caused and will always cause even more tiebreaks. Playing an even number of games will always result in more ties unless you get very very lucky. The best solution is to use the GSL format and make groups a small double elim bracket.
In dota they can use bo2 because they have other scores to use as tiebreakers such as:
time score - often in a group they use time score (how fast you win or slow u lose) kill score - pretty obvious what this is
we don't have that in sc2. We can't use units killed as a score, or workers made or whatever. You can't even use time score because different maps and races cause games to be completely different lengths. You can't compare a TvT on Whirlwind to a ZvP on Neo Planet S, for example.
I think this is a great point and I don't really have a rebuttal. . I still feel like we need to make an improvement somewhere. Perhaps a Bo3 with the 'unimportant games' played off stream that Chairman Ray suggested is best.
MC 3 - 0 Hobbe Whitera 3 - 0 Hobbe Sortof 2 - 1 Whitera Sortof 1 - 2 MC Whitera 2 - 1 MC
So their final scores in this situation:
MC 6 - 3 Whitera 6 - 3
But it is far more likely that whitera doesn't pick up a map against Sortof, as he lost 2-0 he probably isn't going to play his hardest in game 3 and Sortof is the better player anyhow. So it probably looks like this:
MC 3 - 0 Hobbe Whitera 3 - 0 Hobbe Sortof 3 - 0 Whitera Sortof 1 - 2 MC Whitera 2 - 1 MC
So their final scores in this situation:
MC 6 - 3 Whitera 5 - 4
The only time a player will play their hardest in a game 3, when they already lost 2-0 is if they KNOW, and I mean KNOW that winning the 3rd game will keep them in the tourney. The only way to know it will keep them in is for it to be the last series of the group, and they know going in that they need to win it, or atleast take a map, to advance.
I've shown two quite possible variations on your "playing out the whole 3 maps" thing and there are many other possibilities. In most, whitera gets eliminated. What if he loses map 3 to Hobbe but wins map 3 against Sortof? or vice versa? what about if he loses both map 3's? You are creating headaches that don't need to exist, just to stop a rare situation in which Whitera beat MC, they end up with the same series win/loss but MC goes through.
Anyways, are u really trying to argue that Whitera is better than MC? ;p Sorry but if you point is about the best players advancing then Whitera shouldn't even be in the tourney. I love Whitera, hes a great player but he wouldn't even make it past round 1 of GSL code A qualifiers.
I vote that all tourneys just simply move to the GSL/MSL group format, like WCS has done. We never hear arguments about how that is unfair, because its not. You get two bites at the cherry, lose 2 series and you are out. Done.
Edit: There have actually been tourneys where they played out all 3 games, guess what.... spectators didn't like it, players didn't like it, admins didn't like it. No one liked it and it didn't produce better results. Players don't play as hard, it takes A LOT longer to play out the groups, it causes more scheduling issues, players are more tired for the next round. There are many reasons to not force a 3rd game to be played out than there are to require it.
I would be willing to wager a large amount that if we broke out the stats on these situations that the better player advances far more often than not. MC is the better player in this situation, there is no arguing otherwise, both his results on the day and over his career show he is the better player. The better players in a group will lose 2-1 and win 2-0 very often while worse players will usually lose 2-0 and win 2-1.
bo2 will exacerbate the tiebreak issues, i agree with plexa's alternative of just playing all three games of a bo3.
On April 27 2013 10:50 Plexa wrote: If you want to go this way, wouldn't it be better to play all three maps in a bo3? At least in this system you get to always declare a winner, whereas with Bo2's you dont.
On April 27 2013 11:28 emythrel wrote: The only time a player will play their hardest in a game 3, when they already lost 2-0 is if they KNOW, and I mean KNOW that winning the 3rd game will keep them in the tourney. The only way to know it will keep them in is for it to be the last series of the group, and they know going in that they need to win it, or atleast take a map, to advance.
I dont think so. If they know map score counts I don't see why they wouldn't play to win in the third game in the first or second match of a group stage.
Anyways, are u really trying to argue that Whitera is better than MC? ;p Sorry but if you point is about the best players advancing then Whitera shouldn't even be in the tourney. I love Whitera, hes a great player but he wouldn't even make it past round 1 of GSL code A qualifiers.
No. I specifically said a few times that it has nothing to do with the players in the example. I don't think you are really seeing my point. I'm not saying my proposal is the solution. I'm saying there is a major flaw in how we are doing it and at the time of writing the post having Bo2's seemed like the best way to solve it (in a void). It doesn't matter if the frequency of people getting screwed over is low. It's still happening and it does harm the competitive integrity of the competition. I'm sure we could find a way around it that doesn't harm the spectator experience.
Maybe it's a lot of work for not much gain. I'm not sure.
On April 27 2013 11:28 emythrel wrote: The only time a player will play their hardest in a game 3, when they already lost 2-0 is if they KNOW, and I mean KNOW that winning the 3rd game will keep them in the tourney. The only way to know it will keep them in is for it to be the last series of the group, and they know going in that they need to win it, or atleast take a map, to advance.
I dont think so. If they know map score counts I don't see why they wouldn't play to win in the third game in the first or second match of a group stage.
Anyways, are u really trying to argue that Whitera is better than MC? ;p Sorry but if you point is about the best players advancing then Whitera shouldn't even be in the tourney. I love Whitera, hes a great player but he wouldn't even make it past round 1 of GSL code A qualifiers.
No. I specifically said a few times that it has nothing to do with the players in the example. I don't think you are really seeing my point. I'm not saying my proposal is the solution. I'm saying there is a major flaw in how we are doing it and at the time of writing the post having Bo2's seemed like the best way to solve it (in a void). It doesn't matter if the frequency of people getting screwed over is low. It's still happening and it does harm the competitive integrity of the competition. I'm sure we could find a way around it that doesn't harm the spectator experience.
Maybe it's a lot of work for not much gain. I'm not sure.
I agree I dont see your point. Because I dont think you have one. in most situations the best player will advance, no system will ensure the best player advances every time but this one at least gets there more often than not. In this particular situation yes I think Mc should go thru because he lost a bo3 2-1 not 2-0, if you play out the third game you still lost the bo3 2-0 you just then got to play the third game for map score. It isn't a bo3 if you don't win by winning 2 games. Its becomes something else.
So if we say that all games become a 3 game series, not a bo3, you then have to use map score as your main ranking criteria. So whether or not you win against a certain player means nothing. In football leagues you play each team twice, whether or not you beat a specific team doesn't matter because it is game score that matters, not whom you beat. In a bo3 format SERIES wins matter more than individual map scores, its a completely different way of ranking and a completely different mindset.
If you play out all 3 maps you no longer have the need to score by series, therefore beating Whitera 3-0 gives you 3 points instead of 1. You see what I mean? So as in my example above, Whitera would have probably ended up in the same situation. Because both he and MC likely would have 3-0d Hobbe but Whitera likely would have lost 3-0 to Sortof, while MC lost 2-1. So the same situation occurs. The only way things would have changed is if Whitera had won the final game against sortof, which he probably wouldn't have done.
Either way, in all likelyhood, MC advances and Whitera doesn't. Change the players names and the same situation happens. In all likelyhood a good player will lose a 3 game series 2-1 while a lesser player loses to the same player 3-0. All we would be doing is making tourneys last longer, be more complicated, less exciting and more of a headache just to avoid a situation that will come up NO MATTER WHAT.
I can adjust the numbers ten different ways, you still get situations in which Player A beats Player B but then loses to Player C worse than Player B did. A true measure of someone's skill is in how they lose, rather than how they win. Bad players lose outright, good players push their opponents to the limit. There in lies the difference. Whitera MIGHT have been able to take game 3 against Sortof, but since Sortof is the better player (at least imo) its far more likely Whitera loses 3-0 because he is already down 2-0, knows he can't do anything but scrape 1 point and in his head is moving on to the next series, where he can actually gain something significant.
I honestly believe that if all 3 games were played out in every series, we would be sitting here with MC going through and Whitera not. Because MC certainly isn't dropping a map to Hobbe but Whitera might, MC managed to push Whitera to a game 3 the legit way, while Whitera could not push Sortof to one. MC pushed Whitera to a game 3 and lets be honest, if it were any other matchup than PvP, Whitera wouldn't stand a chance against MC, most of the time. Change the players to Taeja and Idra, who is more likely to push their opponents to a game 3 the legit way? Idra or Taeja? We all know its Taeja because he is the better player.
If you get to the GSL final by winning 2-0 the whole way, against the harder opponents, you can easily claim to be better than your opponent who got there 2-1 the whole way to lesser opponents. If you go out of a group by losing 2-1 every game you can claim to be better than the guy who came 4th going 2-0 to everyone. Whitera couldn't push the group winner to a game 3, MC beat the group winner and only lost 2-1 to Whitera. If Whitera beats MC 2-0, we aren't having this conversation. But he didn't. Why should we force players to play out an extra game that they didn't earn? If I beat you 2-0 and then have to play another game against you my first question would be "why?" because I already beat you, you can't win the series and in most situations the 3rd game will make no difference to whom advances. I'd be pretty pissed off if I went out of a tourney because instead of playing bo3's and using map score as a secondary ranking criteria, I was forced to play a game 3 against everyone and get super cheesed by ActionJesus in game 3 and lose, causing one of my opponents to tie with me. Because if map score is your first ranking criteria, head2head score has to be your second, that or playing extra games and in a 2-3 day tourney, no one wants to have to play extra games. The admins don't want it, the players don't want it, the organisers don't want it and I would guess most spectators would rather get to the next round faster too ;p
This is my point, better players get a map score "advantage" because they are better and harder to beat 2-0. So by playing out the 3rd game, no matter what, you are actually making it less likely the better player will advance by giving worse players an extra bite at the cherry for no good reason.
edit:
I don't think I've ever made this many edits and additions to a Tl post before lol. Kept thinking of extra stuff!
I don't understand the argument in the OP. White-rA played less games because he didn't earn a 3rd game by going 0-2. At the same time, MC forced that third game against White-rA in going 1-2. That point is invalid.
There's nothing wrong with the current system, and there is no such thing as bo2. You can't get a best if you tie, and the concept of having 1-1 ties in a set is nonsensical. The right player was eliminated.
On April 27 2013 21:01 esReveR wrote: I don't understand the argument in the OP. White-rA played less games because he didn't earn a 3rd game by going 0-2. At the same time, MC forced that third game against White-rA in going 1-2. That point is invalid.
There's nothing wrong with the current system, and there is no such thing as bo2. You can't get a best if you tie, and the concept of having 1-1 ties in a set is nonsensical. The right player was eliminated.
other competitions do use bo2s, like OGN LoL tourney group stage or chess tournaments that may do like 1 game white 1 game black in a round robin format (dunno if that is used in the actual chess federation type stuff though, but some amateur things can do that)
"bo2" group stage things is just another format, just kind of stylistic. No matter the format, something will come up in tiebreaker like scenarios that makes somebody unhappy. Eventually
On April 27 2013 21:01 esReveR wrote: I don't understand the argument in the OP. White-rA played less games because he didn't earn a 3rd game by going 0-2. At the same time, MC forced that third game against White-rA in going 1-2. That point is invalid.
There's nothing wrong with the current system, and there is no such thing as bo2. You can't get a best if you tie, and the concept of having 1-1 ties in a set is nonsensical. The right player was eliminated.
other competitions do use bo2s, like OGN LoL tourney group stage or chess tournaments that may do like 1 game white 1 game black in a round robin format (dunno if that is used in the actual chess federation type stuff though, but some amateur things can do that)
"bo2" group stage things is just another format, just kind of stylistic. No matter the format, something will come up in tiebreaker like scenarios that makes somebody unhappy. Eventually
I think Bo2 formats tend to be used in games where there are potential map/board imbalances. LoL/Dota do not have perfectly mirrored maps so it can be detrimental for one team to get the "easier" side more often. It doesn't seem like it would work the same way for SC2.
EPT![[image loading]](http://i.imgur.com/VTiWaQw.png)
. I still feel like we need to make an improvement somewhere. Perhaps a Bo3 with the 'unimportant games' played off stream that Chairman Ray suggested is best.
