|
I was working on this post explaining why "Should Team Evil Geniuses Have Established a North American Team house?", was a poorly done article that should not be featured on TL when I started running a few stats on some EG players. As it turned out, both Incontrol and Axslav did worse in the second season of NASL while Machine didn't even qualify for season two.
On the surface, it looks pretty bad. Not only did they do significantly worse, but the general consensus seems to be that Season Two was a weaker field. With the exodus of top Koreans, including 7 of the 15 final tournament participants (excluding open tournament), the field must surely be weaker, right? Seemingly the only saving grace for Season 2 was the addition of a few top end foreigners and foreign-team Koreans such as HerO (14-2), DeMusliM (12-4), PuMa (12-4), and HuK (11-6). But perhaps the more important thing to note is that the bottom of the league also was cut/replaced. After crunching a few numbers, I can safely come to a bold conclusion:
Season Two of NASL had a stronger overall field than Season One.
To come to this conclusion, I compiled the wins and losses for each player that played in Season One and Season Two. A picture of the spreadsheet is spoilered below.
+ Show Spoiler +
Overall, the returning players went 322-278 (54%) in Season One, but only 229-243 (49%) in Season Two. That's a 5% lower winning rate. Same players, same format, less Koreans, yet a worse overall record. Even taking out the aforementioned Axslav and iNcontroL, it still comes out at a 4% drop. Of the 28 returning players, 10 improved their win%, one stayed the same, and 17 did worse.
There will always be variance from game to game so it's not too surprising when one player goes on a very nice/poor run that could skew individual results. However, the combined sample size should be enough to pass mathematical rigor. With near certainty, Season Two has indeed been a tougher regular season than Season One.
So don't fret Starcraft fans. While some could arguably say that the very top of NASL was weakened, the lineup overall was definitely strengthened. The upcoming championship playoffs should be a great exhibition of closely fought games.
|
That's interesting.
Here's my hypothesis that is completely unresearched and unsupported: Koreans in NASL season one preformed less well than foreigner additions because of the timezone and server differences.
So I think you might have something there.
|
On November 30 2011 20:03 SimDawg wrote:
That's interesting.
Here's my hypothesis that is completely unresearched and unsupported: Koreans in NASL season one preformed less well than foreigner additions because of the timezone and server differences.
So I think you might have something there.
Koreans in Season 1 performed really well though. Of the 10 who played in Season 1... only 2 failed to make the finals, if I'm not mistaken (Rainbow and Ace... and Ace was 1 game away from advancing).
|
Honestly, every time somebody makes conclusions about SC2 based on statistics I die a little inside. -_-
This is just one example of how statistics can always be (and often is) wrong.
|
While I agree overall the bottom/middle players of NASL season 2 are probably better than season 1, I wouldn't say for sure that the whole field was overall better. I'd be interested in some more data such as ELO comparisons and such, although I guess that'd be hard as well since most Koreans' foreign TLPD ELO's are a bit off.
|
I hope you're kidding with season 2 having a stronger field 
Even though your statistics might "show" it on these results, there's also another factors that have to be taken into account.
For the players that performed lesser in season 2 than in season 1 you could also take into account all the other matches they played over the same time span. If they also showed a decline in those series it could very well be that some players are just slumping.
Also statistics could very well be influenced by what match-ups these persons had to play.
Therefor saying boldly that season 2 had a stronger field than in season 1 is in my opinion not valid. Maybe a bit more evenly matched, but in season 1 the differences between top and bottem were quite big.
|
If I have player A and B, with A being so much stronger than B that he would win 99 games out of 100 those two played, they would only make a one set (two games) difference in the scoreboard and hence in your statistic.
What I am saying is that some of the Koreans last season were incredibly good, however they are just one player per division!
|
The results don't take into consideration so many factors, some already explained by mithriel, but for instance, many of the returning players participated in tourney's that were alot less established back when NASL 1 was going on and so much more even excluding the "K factor;" what about patch changes and how that influenced the metagame? These factors don't even begin assessing player skill but they already render this evaluation naught.
I for one have seen many of the in-season games and for the most part they are EXTREMELY sub-par, with little to no 'wow-factor' and gives viewers almost NO incentive to tune in until playoffs as the games are pre-recorded and altogether boring.
Maybe wait 3-4 seasons (if it lasts that long) to crunch numbers and see if leaving out the best players in the world in your "global" SC2 tournament really does inspire some random influx of skill. I bet my money that it doesn't.
|
Stronger yes. The average player ability has risen in the sense of most of the garbage being removed. This has definitely resulted in more even games.
Eg. Preslump MC was slightly better than Sen who is much much better than Incontrol. The gap between the top players are narrower than the differences around the bottom.
It is interesting to note that the "garbage" who remained, NASL managed to retain qxc, hasu and haypro, whose recent improvement and resurgence in other competitions truly underlines their improved score in NASL, which goes against the trend. While many can argue against their inclusion before s2, few can after their results of the past few weeks. A lucky fluke for NASL indeed.
|
Season 1 had a lot of deadweight players, who were essentially free wins. Also groups were 10 in size instead of 8.
Now season 2 had no koreans, but the deadweight players were also gone. So that's how you got your 50ish%->50% statistic.
Championship bracket is not going to be as good without the top tier players.
|
It depends on how you define a strong field. The top is worse, the depth is probably stronger. How you way these two against each other is arguable.
|
That is a bold conclusion indeed. The saying, "Lies, damned lies, and statistics" kind of comes to mind here.
I’m no expert in statistics, but I would say that this exercise looks somewhat meaningless. 5% drop or increase is so small that it can be the result of players just having one or two “off days”. It may also depend other intangibles, such as did the players focus on practicing for these games or did they focus on others such as all the MLGs/IPLs/DreamHacks out there? What match-ups where played? Did the Koreans really care that much about this tournament? Etc.
All in all one can try to judge the competition by looking at the returning players but I believe it should be done after the season and by looking at their final placements. It is still not really a fair comparison due to the same reasons that others point out for the exercise you did but I believe it to be slightly more fair.
|
so many walkover wins dropped out
thats why players are winning less
|
I enjoyed season2 way more.
When I see 2 koreans playing I couldn't care less who wins, there is no storyline for me.
But with 'foreigners' it's far more enjoyable too watch, and tbh the level of foreigners is getting really close to korea.
Such tournaments only help to improve that because the players can get more motivated.
They can fight koreans at the other 5milllion lans out there.
edit: Well I guess the chances of Puma or Hero winning are still pretty damn high ^^
|
Also from SoTG and what Tyler/Incontrol/Sheth said about NASL. The format was completely different from Season 1 to Season 2 due to scheduling issues. So in Season 1 where you had a player play one match a week, Season 2 they said they played several matches a day. So if you had a bad day in Season 1 you could hopefully bounce back next week with just one loss. This time around you had a bad day and you had 3-4 losses. Take into account being able to prepare for each opponent each week if you were so inclined(Season 1) and playing multiple matches a day(Season 2) left less room for that.
Completely different beasts AKA Season 1 and Season 2 are just not comparable.
|
Don't forget how long ago those matches were played...the whole NASL online season was probably long over by the time Axslav really hit the level of one of NA's best, beating Koreans and such. Obviously this is speculation but the team house would have only recently opened when a decent number of those matches were played.
|
That is a shady statistic. The bottom of the league is better in NASL 2, but the top is most likely not significantly better., if not worse.
1. Walkovers skew stats
Division A
Not included. Reason stated in part 2.
Division B
July w/o Whitera (Whitera -2, July +2)
Sjow w/o Ensnare (Ensnare -2, Sjow +2)
Whitera w/o Ensnare (Ensnare -4, Whitera +2)
Whitera - 13-8 in S1. 11-6 with w/o results removed. From 61.9% to 64.7%. (29% in S2)
Sjow - 13-8 in S1. 11-8 with w/o results removed. From 62% to 57.9%. (47% in S2) - (If Sjow had lost the w/o removed match he would have been 11-10 and 52.4% whichi s much closer to how he did in S2)
Division B break down. Sjow moves closer to his S2 %. Whitera's difference remains significantly different.
Division C
Strelok w/o Naniwa (Naniwa - 0, Strelok +2)
Naniwa w/o Moonglade (Moonglade -2, Naniwa +2)
Strelok - 17-6 in S1. 15-6 with w/o results removed. From 73.9% to 71.4%. (67% in S2)
Moonglade - 9-13 in S1. 9-11 with w/o results removed. 40.9% to 45%. (47% in S2)
Division C breakdown. Both players go closer to their S2 %s with w/os removed.
Division D
Bratok w/o MC (MC -0, Bratok +2)
Select w/o Bratok (Bratok -2, Select +2)
MC w/o Goody (Goody -2, MC +2)
Bratok - 10-12 in S1. 8-10 with w/o results removed. From 45% to 44.4%. (67% in S2)
Select - 16-5 in S1. 14-5 with w/o results removed. From 76.2% to 73.7%. (56% in S2)
Division D breakdown. To be honest, there is not much difference in this group with w/os removed.
Division E
Not included. Reason stated in Part 2.
2. Artosis/Painuser.
To an extent, they were free wins throughout the season. Whether they played the matches out of whether they provided walkovers.
These 2 players (I won't argue whether they should have been in the season from the start) skewed the stats favorably for a lot of players due to how bad they were in the season.
In division A, you could argue 6 players on your list got 2 free wins.
If you take Artosis out of the group...
Sheth from 14-6 to 12-6. 70% to 66.6%.(59% in S2)
Morrow from 15-8 to 13-8. 65% to 61.9%. (71% in S2)
Kiwikaki from 13-6 to 11-6. 68% to 64.7%. (53% in S2)
Fenix from 13-8 to 11-8. 62% to 57.9%. (31% in S2) Fenix was actually 41.7% if you ignore the w/os.
TLO from 10-11 to 8-11. 48% to 42.1%. (33% in S2).
Vibe from 8-12 to 6-12. 40% to 33.3%%. (39% in S2)
Only names which taking these games out takes it further away from the S2 % is Morrow and Vibe. Fenix still has a significant difference.
In division E, you could argue 6 players on your list got 2 free wins.
If you take PainUser out of the group...
Sen from 16-5 to 14-5. 76% to 73.7%. (71% in S2)
Idra from 13-8 to 11-8. 62% to 57.9%. (71% in S2)
Mana from 11-9 to 9-9. 55% to 50%. (60% in S2)
Socke from 10-12 to 8-11. 45% to 42.1%.(44% in S2)
Tyler from 8-13 to 6-13. 38% to 31.6%. (33% in S2)
Cloud from 8-14 to 6-13. 36% to 31.6%. (47% in S2)
Only names which taking these games out takes it further away from the S2 % is Idra, Mana, and Cloud. (Socke difference is marginal)
Overall with w/os and Artosis/PU results removed.
322-278 S1 - 54%
288-274 S1 ADJ - 51%
229-243 S2 - 49%
There isn't really a significant difference.
They just took the better players out of S1 and replaced them with players better than the returning ones on this list.
This doesn't mean that the replacing players are better than the ones they replaced. Just that they remain better than those on this list.
Also, some of the free wins towards the bottom were eliminated which shifted the weaker players to the bottom. (Though you could argue Softball was added as a free win.)
(I know this isn't exactly a clear cut statistic showing any significant proof. Of course, such a statistic cannot be done due to variance of players, balance changes, and other such changes. I just had some free time and wanted to show a possible explanation for the supposed difference in win %. 2% change in win rate would be insignificant.)
|
On November 30 2011 22:55 Corrik wrote:
2. Artosis/Painuser.
To an extent, they were free wins throughout the season. Whether they played the matches out of whether they provided walkovers.
You mean whether they just left the player hanging, never appearing to the scheduled match in the first place, which lead to their Korean opponents getting up at 4am just to wait for several hours before getting the defwin?
|
On November 30 2011 22:59 JustPassingBy wrote:Show nested quote +On November 30 2011 22:55 Corrik wrote:
2. Artosis/Painuser.
To an extent, they were free wins throughout the season. Whether they played the matches out of whether they provided walkovers. You mean whether they just left the player hanging, never appearing to the scheduled match in the first place, which lead to their Korean opponents getting up at 4am just to wait for several hours before getting the defwin?
Yep. That's exactly what I mean.
|
You can't just compare season 1 and season 2 stats of the returning players. You have to have factor in the non-NASL stats of these players over that time period as well. Like as not, many of the invited players in season 1 have been surpassed by the new talent that qualified in season 2, but such players were probably still newcomers to the scene in the time period around season 1. On an absolute level, the skill level has most likely increased in season 2, but the relative skill level of the competition to the skill level of the whole body of SC2 progamers is most likely worse.
|
|
|
|
|
|
EPT![[image loading]](http://starcraft.sageground.com/images/NASLS1vS2.jpg)
