Flipsid3 Tactics SC2 Academy: Initial Roster and Postmortem
In early February, I posted about the opening of the Flipsid3 Tactics StarCraft II Academy team on Reddit, Teamliquid, and our website. Having been apart of the recruitment process for Clarity Gaming’s Academy, I knew that providing a competitive and informative tryout environment could be beneficial not only for the organization in finding strong new players, but also for players looking to make the jump into playing the game at a high(er) level. With that thought in mind, Kevin “Paperboat” Cosgrove and I went about designing a system that could help us accurately find 8-9 players to build our new roster.
Before I get into the details of the tryout, I’d like to first congratulate and announce the players who survived the six-week tryout. They are all truly outstanding players, and I can’t wait to see how they progress in the future.
Joseph “Honeybear” Millard (Z)
Aaron “Coldscars” Eich (T)
Enes “Reprieve” Apelqvist (T)
Francis “friedpork” Chik (Z)
Kamil “kel” Malinowski (Z)
Brady “Rengen” Cassada (P)
Tan-AnNam “Nexa” Nguyen (P)
Jacob “picur” Molyneux (Z)
João “MaLeS” Oliveira (P)
Team Captain: Kevin “Paperboat” Cosgrove (Z)
Aaron “Coldscars” Eich (T)
Enes “Reprieve” Apelqvist (T)
Francis “friedpork” Chik (Z)
Kamil “kel” Malinowski (Z)
Brady “Rengen” Cassada (P)
Tan-AnNam “Nexa” Nguyen (P)
Jacob “picur” Molyneux (Z)
João “MaLeS” Oliveira (P)
Team Captain: Kevin “Paperboat” Cosgrove (Z)
In the Beginning
When I decided to launch the academy, I knew the first people I’d ask for help were Michael “J” “Cav” Bramley and Kevin “Paperboat” Cosgrove. Mike was one of the guys who ran the academy squad at Clarity Gaming, and had observed the tryout process from up top. Kevin was a member of Clarity Gaming for a good amount of time, and was later asked to head the reopened academy (which subsequently ended following Clarity Gaming’s closure three months later).
What we agreed on is we wanted some sort of automatically updating leaderboard, one which we could track progress on. The real question was simply: what should we be tracking? Ladder rank? In-house games? How can we track in-house games with a projected tryout size of 50-100 people?
The In-House Ranking System
I proposed using the Elo system to track in-house ranking games. After some discussion, we decided that assigning each player five matches per week, for a total of 30 over the tryout, would be both a sufficient sample size to determine who the best players in the tryout were, as well as manageable for the players. Agreeing that this was a satisfactory starting point for building the tryout, we went to work on building a match assigner, as well as a way to enter and store matches and the player's’ Elo rating.
From a perspective of system functionality, the in-house ranking system worked great. Kevin, a third year computer science major at Vassar, did most of the heavy work. He build the match scheduler from hand in C#, as well as setting up the Elo calculator in Python + JS (in conjunction with Google Sheets API). Big props to whoever’s Elo.py we ripped from github for this project. I made the very ugly but functioning GUI in Python. We tested the system somewhat extensively and determined it was ready to roll.
This is what the match reporter looks like. Straight outta 1998.
I’ll explain very briefly how the system works. Players get assigned five different opponents from the tryout every week. They are responsible for scheduling their matches with their opponent. The format is Best-of-2. Many players were confused why we used Bo2, but to keep it brief: a player with a lower rating gets rewarded for playing a higher rating player to a draw. Grubby’s VLOG from Korea has great insight as to why Korean team houses use Bo2. After the players play their two matches, they are inputted into the match reporter program. The program then uses the Elo algorithm to determine their new rating. Everyone started at 1200, so if you've played on chess.com or some other site with ratings, it may be easier to visualize. The ratings are then stored and accessible via a Google Sheet. It was a leaderboard that could be updated and viewed in real-time, as we inputted the matches.
The sheet itself that tracked the in-house matches.
I received about 65 applicants for the tryout during the official filing time, which we were very thrilled about. I gathered them all into one Skype group, and we finalized official details regarding tryout procedures. With everyone amped up and ready to roll, we started week 1 on February 22nd. It was both simultaneously awesome and nightmarish, all wrapped into one.
I thought I had made the initial tryout post very clear that this would be an intense tryout which required ample participation from everyone. As soon as the first week’s matches were released (remember, each player only had 5 matches to play against other players in the tryout group, and could be scheduled at any time), we had five or six players instantly say, “hey, thanks but this is not really for me.” I wasn’t really too mad, but it really wonked out the system when people didn’t play their matches.
By the end of the first week, we had lost 20 or so players, all of which had forfeited their matches. It left the Elo leaderboard in essential disarray. So many players were behind on matches, and their rating was either higher or lower than those who had been able to play all five of their assigned matches. After reviewing the situation, I decided to award make-up matches to those who had forfeiting opponents. I promised that everyone would play 30 matches by the end of the six weeks.
Unfortunately, it only continued during the second week. More and more people quit, and it left people missing a great deal of their matches. Though the strongest players still generally rose to the top, it must not have been fun for players who only had played 4/10 of their matches because of forfeits. I cancelled in-house matches after the second week. Players seemed to have a lot of difficulty scheduling their matches, which is understandable. Many players either work or go to school, and there was a solid amount of diversity in location. A guy with a full-time job in Canada understandably would have trouble finding a common time to play the 16 year-old student from Sweden. As such, we decided it was a system that just wasn’t practical given the circumstances. It was still a good tool for honing in on potential players for the final roster. I’m overall really glad we went for it.
What I Learned from Hosting Tournaments
In developing our plan for the academy, we decided to run four tournaments open only to those who had applied in February, followed by a single tournament open to those who hadn’t. Each one gave a direct invite to the team for first place. These were relatively consistent, as they had a set time every Saturday and it was either “show up or don’t, whatever yo.” I still learned a lot about being a tournament admin during these events, though.
There were some serious tensions during the first tournament. I work on weekends, so I was not available to help run this one. I left it to Paperboat to admin the whole thing, but my ruleset was rather bad and was missing some very important information. The most important, which I received many messages about on my phone, was that of server selection. Two players had a very heated disagreement about which server to host on when it was a US player vs EU player. I was able to explain and settle the issue, but that match had already been played.
The second issue was I had made the tournament double elimination, and decided the finals should be two Bo5s if the loser wins the first Bo5. With how late the tournament went, and with one player having to play the final bo5 at like 4:00 AM CET, it was a bit of a disaster. Following this tournament, I went back and made sure the ruleset was very easy to follow, and the rest of the tournaments went great. I was very happy with how everyone played, and I got to see some really great games.
The open tournament was a pretty horrible experience for me. I decided to make the groups round robin, and everyone seemed pretty happy about getting to play more matches. That was until they ran into a 3-way tiebreaker. Those were pretty difficult to explain and some of them just outright didn’t make sense. If you remember the Dreamhack tournament where advancements were literally decided by drawing names, it felt like that. By the end of the tournament though, there was a fantastic finals between Jamileon and MaLeS. They are both super cool guys, and I was very happy that one of them was going to end up on the final team.
Picking the Team
Originally, we had advertised the roster selection as the following:
(4) Players from the in-house tournaments
(2) Players from the top of the in-house ranking sheet at the end of the six weeks
(1) Player from the open tournament
(1-2) Players that we hand pick
(2) Players from the top of the in-house ranking sheet at the end of the six weeks
(1) Player from the open tournament
(1-2) Players that we hand pick
With the abolition of the in-house rankings after the second week, we added those two spots from the top 2 leaderboard picks to our staff picks. So in the end, we had five players qualify directly from winning a tournament, and the last four were selected after some deliberation. It was unfortunate that we were not able to deliver on what we had promised the players originally, but most agreed that this was a better way to finalize and select the last four members.
TL;DR and What I Learned
A six-week tryout was insane. If we had kept the pace we had set the first two weeks, I imagine a lot more players would have dropped out because of how intense the system was. I think I was way too idealistic about the time commitment of most players. There were a good amount who really gave it their all and worked their asses off to schedule and play all their games as well as the tournaments, and I’m extremely grateful and impressed by their efforts. In the end, I was disappointed the system didn’t work out, and I put a lot of blame on the flaky players who had signed up originally. I understand though that it was a really big commitment, something I may not have made expressly clear at the start.
That said, I’m really glad we did this six-week tryout. I met a lot of really cool players, and I am confident that all of them can really find a place at the top of the scene if they keep working at it. A lot of players told me they had a lot of fun and learned a lot, which was what I hoped for when we created this monster.
I learned a ton about tournament and league administration. Though I didn't make any huge, fatal errors, there were certainly some hiccups along the way. I feel very confident that if I ever try to do something so extensive, it will be executed better next time. That's not to say I think I did a poor job, but there's always room for improvement.
Most importantly, we built a brand new team which I'm confident will do great things! I'm looking forward to running some in-house practice with them, as well as playing in team leagues and various events. This is only the beginning of Flipsid3's venture into amateur StarCraft II. We're here to make a difference in the ecosystem.