Ever since the You Are The Golf King MSL in 2004, the two most prestigious SC:BW tournaments in Korea have run Bo5 formats for their finals, instead of the Bo7 popular among SC2 tourney organizers, most notably GSL. In fact, the Bo5 has been the default finals setup for pretty much every Brood War tournament organizer. As mentioned, the MSL tinkered around with a Bo7 but quickly switched back.
SC2 tournament organizers should follow their predecessors' example and switch to a format involving only Bo5 series. Bo7 series are very good at awarding the win to the better player, but they are not so good at delivering an entertaining experience.
The Last Game of a Finals is Always the Best One
This principle was laid out by Arrian in his excellent article, A Finale in Five. When he said that the final game of the series was always the best, he was referring to the game 5 in a Bo5 series, or in the GSL's case, a game 7 in a Bo7 series. Those are games when everything is on the line. Whoever wins that one game wins the whole series.
the last game really is always the best
It's easy to show that this principle is true, since we only have a small number of series to look at. There are six Bo7 series in the history of big-event competitive SC2 that have made it to the 7th and final game: ThorZaIN vs NaNiwa in TSL 3, DRG vs MMA in the 2011 Blizzard Cup, MKP vs NesTea in the second GSL Open, MVP vs Squirtle in the 2012 GSL Season 2, PuMa vs MC in NASL Season 1, and HerO vs PuMa at the last Dreamhack Winter. Six series in the entire history of SC2. Six out of forty-five (13%), actually. I counted.
In three of those six series, the seventh game was definitely considered to be the best. You just have to look at the recommended game polls for Naniwa vs Thorzain, DRG vs MMA, and HerO vs Puma to see the truth of that.
naniwa vs thorzain game 7
But the two series that I want to focus on are MKP vs Nestea and Mvp vs Squirtle. Both those series ended with games that, if viewed in a vacuum, would have been considered terrible. Somehow, though, they were highly rated by spectators as amazing games. That's the power of a game 7.
In game 7 of MKP vs Nestea, MKP went for an all-in 2 rax play, a build identical to the one he had performed in game 1. In fact, games 1 and 7 were almost identical in the way they played out -- so why did Lovedrop, in writing his finals recap article, rate game 1 much lower than he did game 7? And why did I, while watching the series live, feel many times more excited watching game 7 than game 1? It was the all-or-nothing stakes which hung over the game that bumped it up a few notches in intensity. The fact that it was the final game in the series made it the best.
In game 7 of Mvp vs Squirtle, Mvp went for an early rax all-in. Both players made mistakes, and the game ended quickly. It wasn't a good game at all! Yet in the recommended games poll, the game received a five-star rating from 70% of TLers. The fact that it was the final game in the series increased its perceived quality.
This isn't a new phenomenon. Game 5s in Brood War had a similar aura about them. One that stands out in my mind in particular was the final game of the exhibition series between Bisu and Jaedong after GOM season 2. Game 5 involved Jaedong running zerglings up a ramp on the map Andromeda and finishing the game rather quickly after that. The rest of the series featured some unusually high quality PvZ, but for the life of me, I can only remember details of the game that lasted eight minutes!
In Brood War, the frequency of finals that went to game 5 was 30% (15/50 -- yes, I counted again), substantially higher than the 13% for Bo7s in SC2. It makes sense that that's the case. Say you have two players, one who completely outclasses the other. Let's say the chance of Player A beating Player B is about 90%. In a Bo3, Player B actually has an 18% chance of taking the series to a game 3 (and an additional 1% chance of winning before then). But the chance of taking a Bo5 to game 5 is much lower -- about 5% -- and the chance of taking a Bo7 to game 7 is lower still (less than 2%). It's skill difference that makes a series more likely to go to the final game in a Bo5 than a Bo7.
To sum up: the final game of the series is (almost) always the best, and a Bo5 format leads to more final games than a Bo7 format.
Anticlimax
Often, after a Bo7 reaches the point of 3-0 or 3-1, things can get rather boring, despite the fact that a friggin' finals is being played. The result of the series feels like a foregone conclusion. The spectators hit alt-tab, and other important internet things are done while the SC2 match finishes in the background.
If the match feels like a foregone conclusion, that's because it pretty much is. There have been 31 Bo7 series that reached the point of 3-1 or 3-0. Of those, in only one match did the disadvantaged player eventually win the series.
That's right. You could have closed out the stream and gone to bed after a 3-1 or 3-0 result and only regretted it 3.2% of the time.
This feeling of anti-climax is eliminated by a Bo5 finals format. Even players in an 0-2 hole don't seem incapable of pulling the match out.
Maps
Maps are some of the most important parts of a pro scene. Just ask progamers who were competing when Scrap Station and Steppes of War were still floating around.
A smaller series allows players to eliminate unfair maps from the map pool via veto. It allows progamers to practice for fewer maps when preparing for the finals, condensing their practice time into prepping five map-specific strategies instead of seven. Watch Jjakji vs Leenock game 1 for an example of how a prepared strategy (the hidden command center behind the mineral patches) can change a game and improve the quality of a series.
The NFL vs the NBA
The NFL and the NBA are two institutions that take an entirely different approach to the playoffs. The NBA desperately wants the best team to win the finals, period -- from the start of the playoffs, teams play Bo7 series to ensure that the better team advances. The NFL, however, plays only one game per round. Winner takes all. This leads to more unpredictable playoffs. It's rare that the "best team" going into the playoffs wins the Super Bowl.
Yet that's not a bad thing. It creates an incredibly exciting postseason every year. SC2 organizations need not be hung up on the idea that the best player must win the tournament. There's a lot of luck associated with tournaments, anyway... the best players regularly get knocked out early on due to cheese or other flukes (just ask Bisu). That's part of what makes ESPORTS, and sports in general, so exciting. Sometimes the best player doesn't win.
incredibly insightful posts. i think the progression of boX give the viewers an interesting experience. watching the TSL qualies was a lot of fun, especially watching Ret's stream in the bo1 rounds. wonder what it would look like if some sc2 tournament organizer ran an entire tourney of bo1.
Your portion on anticlimax (3.2% of games resulted in a reversal at the brink) was interesting, more so because it is another example of how the GSTL refuses common knowledge or regular facts.
I grew up with Bo7 so it's hard for me to see it your way. At best, I like both ways. Most of the time I'm more concerned about whether silly double elimination-losers bracket-wore a brown shirt with black pants-didn't tip the boothgirl, inane type rules that have notably affected the finals of MLG beforehand.
Aside from my desire that any final should start clean, 0 to 0, no matter the earlier circumstances, I don't care all too much about Bo5 or Bo7. SC2 could use both. I would enjoy it either way. Thank you for taking the time to write this out though, it was interesting and made me think about the matter.
I was glued to my screen with my heart beating like hell during Mvp v Squirtle. I think I would being willing to trade a few games to get that feeling more.
Also,
On July 16 2012 15:42 Probe1 wrote: Aside from my desire that any final should start clean, 0 to 0, no matter the earlier circumstances
Bo5 > Bo7 in my books, from a viewer perspective. Feels like in a Bo7, the series just takes way too long, and the games aren't necessarily better. In fact, I can totally see someone half-assing a map they're not comfortable on with some lame strategy carrying the mindset of "Oh, if it works great, if not, I have 3 more losses till I lose". Bo5 comebacks are more frequent, more exciting, and it feels as if each game counts more because each individual game counts as a higher percentage of the overall goal.
Btw, GGPlay/Iris Game 5 finals #1. Shit was insaneeeeeeeee.
I definitely prefer the Bo5 system better, its something that I got really used to in BW
As for the excitement of a final game in a series, I feel that excitement has yet to completely pass over to other eSports, namely LoL; the only time I remember being excited for a final LoL game was during the most recent Dreamhack tourney.
Didn't MKP vs DRG in the last MLG they faced in the finals (winter arena or spring arena?) also go to game 7? Either way yeah I don't mind either though personally
SC2 tournaments started using Bo7 because in the early days SC2 games were usually extremely short (Steppes of war etc..) which led to silly things like the finals being over in less than 30mn of gametime. Now that things are stabilizing a bit, going back to bo5 could indeed be a good idea.
That said I agree with Probe1, we need to get rid of double elimination brackets, sure it creates more games but in the end the finals are just boring since one of the two players starts with a huge advantage. Single elimination brackets are way more exciting to watch because every game the players tournament life is on the line. Double elimination brackets on the other create these feelings of "oh well it's no big deal he lost since there's always loser bracket" and "oh well loser bracket doesnt really matter, the guy coming from the winner bracket will win the finals anyway" which kills excitement like nothing else. And I won't even talk about extended series :<
So in the end yeah, bo5 over bo7? why not, but double elim and extended series are a bigger priority imo.
On July 16 2012 15:27 motbob wrote: It's rare that the "best team" going into the playoffs wins the Super Bowl.
Yet that's not a bad thing. It creates an incredibly exciting postseason every year. SC2 organizations need not be hung up on the idea that the best player must win the tournament. There's a lot of luck associated with tournaments, anyway... the best players regularly get knocked out early on due to cheese or other flukes (just ask Bisu). That's part of what makes ESPORTS, and sports in general, so exciting. Sometimes the best player doesn't win.
I really really hate this approach and somehow this is only a problem for team sports. In tennis for example, in a high percentage of the time, the best player actually wins.
It's a good read, but Bo7 is IMO the best way to go. The better player will win and the excitement from the last game of a Bo7 is usually much better than the excitement that comes from the last game of a Bo5 (see games like Moon vs Huk, and countless IEMs, a lot of GSL playoff games which, aside from sC vs Nestea don't seem to be as exciting as a Bo7 final).
And there needs to be something which separates a Grand Final from a regular play off match and I feel Bo7 does that.
Also Leenock Jjakji ended 4-2, but the last game was nail-bitingly close and intense.
I don't know. You've made good points, but I feel just for the better excitement and story that comes with a full-series Bo7, a Bo7 is much better than a Bo5.
From a viewer perspective, bo5 is clearly superior to bo7. From a player perspective, it's always better with more games since it gives more room for skill and less for luck. One could discuss how big the difference is between a bo5 and a bo7 though, I think most people agree that if a player wins a bo5, they were the better player except in extreme cases.
I believe the biggest problem with Bo5 in SC2 is that the games are just too short. I believe a standard 4 games Bo5 in SC2 takes probably as long as the 3-0 OSL-Finals ZvZ blowout between Jaedong-Yarnc. And a 5 games series in BW takes longer than any Bo7 SC2 series every played, so there is enough build up for the final confrontation. I often wish that SC2 would just play out 5% slower, that's a thought I have had for a long time actually. So I still like Bo7 for the Grand Finals, basically because the games are so fast I need me some more games.
I applaud you good sir. This is something I've thought about quite a bit and I also prefer the bo5. I thought I was just so used to it from BW but I find bo7 often quite boring and you made a convincing case of why that is. Especially the 3-0 situation is depressing. The series drags on but everybody knows how this is gonna end.
The problem as people have mentioned is that SC2 games can end much more quickly, and Bo5 series can be over in a flash. The point about anticlimax is extremely astute as well--going down 3-0 or 3-1 just seems like an insurmountable problem many times. The NFL vs NBA point was also very interesting. I indeed think Bo5 is the way to go from a viewer's perspective. However, from a player's perspective in SC2, Bo7 definitely will be their favored format because they want to know that their hard practice and more consistent play will be rewarded etc I think.
The "better player doesn't always win" is a bit iffy though, since applied to SC2, it really sucks to have it decided with a comparatively shortened series.
It's easy to show that this principle is true, since we only have a small number of series to look at. There are six Bo7 series in the history of big-event competitive SC2 that have made it to the 7th and final game: ThorZaIN vs NaNiwa in TSL 3, DRG vs MMA in the 2011 Blizzard Cup, MKP vs NesTea in the second GSL Open, MVP vs Squirtle in the 2012 GSL Season 2, PuMa vs MC in NASL Season 1, and HerO vs PuMa at the last Dreamhack Winter. Six series in the entire history of SC2. Six out of forty-five (13%), actually. I counted.
That fact that the sample size is small isnt helping Your case, its irresponsible to draw conclusion from the small amount of data.
PS. I personaly like bo7 system. PS2. Actually after giving it some thought, i think that entire point is tottaly moot, the most important factor when considering bo7 vs bo5 is time consumption and ability to properly forsee and shedule the event. The organizers need to fill the time and stay at shedule at the same time, they will choose the format based on this rather than producing best game possible. People remmeber events based on production value, funny things happening and they favorite making, gameplay value as long isnt reaaaaaaly low doesnt matter that match (not to mention that best game ever for someone might be crappy for someone else).
People tend to overestimate the effect of extending the serie in term of helping the best player anyway. I think your point is especially true for a prepared tournament like GSL/OSL, as Bo5 ensures every map will be thouroughly prepared. In the same vein, I think loser pick map should disapear. And as a whole, use less map for a given tournament, but rotate them more. Also it's funny how SC2 is not less volatile than BW according to some people (when it comes to bw vs sc2 discussion), and when it comes to choosing a tournament format, you need Bo11 from the Ro64 on or it's unfair to the players. I think it has to do with how much people hate cheese in SC2, while it was often well received in BW. I'd therefore blame Artosis.
You know, there are some sports -like american football, football, rugby- where players simply need to have some days of rest after games for various reasons. If the NFL could have games every 2nd day, they surely would have bo3 playoffs at least. But as there have to be pauses between the games, with the current playoff system, going to bo3 would suddenly make playoffs lasting up to 12 weeks instead of 4. That's right, the super Bowl teams would have 3 less months of offseason compared to teams who didn't make the playoffs.
It not that NFL, or the World Cup, think bo1 is the best and fairest system. It is just the only one practically applicable for them.
And to the point of maps: If you allow maps being played twice in a bo7, this could very well increase to power of preparation. But I agree bo5 is optimal for sc:bw and sc2.
It's easy to show that this principle is true, since we only have a small number of series to look at. There are six Bo7 series in the history of big-event competitive SC2 that have made it to the 7th and final game: ThorZaIN vs NaNiwa in TSL 3, DRG vs MMA in the 2011 Blizzard Cup, MKP vs NesTea in the second GSL Open, MVP vs Squirtle in the 2012 GSL Season 2, PuMa vs MC in NASL Season 1, and HerO vs PuMa at the last Dreamhack Winter. Six series in the entire history of SC2. Six out of forty-five (13%), actually. I counted.
That fact that the sample size is small isnt helping Your case, its irresponsible to draw conclusion from the small amount of data.
45 series is actually a very decent sample size for this kind of data and will net you decent results. People here sometimes tend to way overestimate the numbers required to reach statistical conclusions. The important point he is making is a) that it happens not often (which you don't even need data for... 2 people being so equal that they need 7 games to decide who is better is obviously something that happens very rarely) and b) that it happens less times than it would in a Bo5 situation. Which is also easily to prove intuitively and statistically.
edit: So the only thing he really is looking for is the 13% game7 in Bo7s is (statistically) significantly lower than the 30% game5 in Bo5s, and there is no doubt in my mind that it is. What the data actually cannot really show (and in this way your criticism would be valid) is if 13% would still be the number of game7 in a sample of like 200 Bo7s. This would be very unlikely, with 45 cases the number could very well range somewhere in between 10-20%, as one more series would already shift the results by a few percentage points.
EPT
