EPTNow that GSTL has started, it's safe for all of us to share our thoughts on putting together a Fantasy GSTL team and maybe put out theories to the test. I'll share my team as well at the end as well as why I chose them. Think of this as a normal team discussion post but with some general theorizing beforehand.
Terms:
FGSTL-Fantasy GSTL
Cost-Initial numbers assigned to players for filling the 36/13 team size in the beginning of the season
Points-The score players accumulate for the fantasy teams
Player-team points-The points a player gets for being on the winning team
Owner- TL members making Fantasy teams
Fteam-Fantasy team (compared to actual RL pro-teams)
Match-Best of 7 between two teams
Game-Individual games of two players from both teams
PICKING YOUR TEAM
PLAYER COST
At the very heart of the league is the cost of the players. Sure, there are the points that they can earn as well as their relative roles in the team (Starter, Ace, Sniper, etc.), but even more important than all of that is the cost of that player to put them into your team. I don't exactly know how TL chooses the various distribution of costs, but I'm pretty sure they did a really good job with it, seeing as how there aren't any glaringly obvious mistakes.
In theory, if TL made the costs perfectly for each player, then no one would win FGSTL. Or more accurately, everyone who made a full 36/36, 13/13 team would have exactly the same amount of points because the cost/point ratio for the entire season for each player would be exactly the same. Note that this precludes trading, which I will also deal with later. For example, that would mean MMA, with a cost of 10 points would return 20 points, then Clide, with a cost of 4 would return 8 points. Thus everyone would have 46 points by the end of the season ([36-13]*2).
Therefore, the goal of picking your team to maximize your points (again, disregarding trading), it to figure out not who is going to do well, but where TL has made a mistake in determining the cost. Your goal is to figure out which players are actually worth more points than their cost would dictate and the opposite for the anti-team. There are a few more subtleties with anti-team, but all in good time. So in a separate example, MC might earn 40 points, a huge amount with his cost of 9. However, if an unknown like Hero earns 5 points with his cost of 1, he is a better player because TL has made a mistake in assigning either Hero a lower cost than he is worth or MC a high cost than he is worth. This is also due to the fact that the cost works on a integer system so there isn't much room for fine tuning. Though it remains to be seen how close the final FGSTL team scores will be, fortunes could be made in the decimals.
It's for this reason that I am afraid of choosing players like MMA, DRG, MC, Nestea, etc. because while they are almost guaranteed to do fairly well, it's not a question of how well they do, but how well they do based on their initial cost. For people picking MMA, last night is an indication of that. MMA dropped a potential 5 point pickup, and was instead worth 0 points last night. For everyone picking him, it was a pretty huge blow to have 25% of your investment depreciate in trade value by 5.8% in a single game and possible cause more importantly, the opportunity cost of picking up another big player. Sure, he has another 4 games to play, but as I said early, the point isn't picking who is going to do well, it's who is going to do well relative to their cost, and MMA has already lost 20% of his chance to score points.
And while DRG owners are rejoicing, if it ends up he never plays another game, or doesn't do quite as well in the remaining games, there is still a possibility that he will under perform for his cost, leading to a end season ratio loss for those owners.
My main point here is, don't stress over who will do well in GSTL, but instead think about whether that player will do well enough to justify his cost, something that is especially relevant for such higher cost players.
NON-SKILL DEPENDENT VARIABLES
While the actual skill of the player is critical in determining their potential point gain, in a team format, it's almost less important who is good, and to determine who is useful. There are also a host of other factors that do not depend on the actual skill of the players in question.
One important thing to note about players and them gaining points is what teams they play and the teams they play on.
Firstly, GSTL is interesting in that it is divided into two halves that do not interact. Many people have noticed that Jupiter is significantly more competitive than Saturn, which causes different effects on teams in both groups.
Because of the more competitive teams in Jupiter, matches will more likely go on for more games, meaning that aces will come into effect more often. As we saw last night, the two aces, DRG and MMA were both played, something that would not happen if the match only went for 4-6 games. Something else of note: in every match, there will always be an ace that gets played. This is because ultimately, at least one team will always come down to their last player. In Saturn, where more one-sided matches are predicted, the probability of two aces coming out in the same match is less likely.
Because of the schedules, there are a few teams to take special note of: TSL, F.United, IM, ST. You might notice that these are the two weakest teams of the "strong" group and the two strongest teams of the "weak" group. Because you also get +1 point for players on the winning teams regardless of playtime, it causes a slight bias to choose IM and ST players while avoiding TSL and F.United (I'd write FU, but that just seems unfortunate for them). This shouldn't be the deciding factor, but if it ever comes down to a perfect tie between two players, use this to break that tie.
Imagine also the player Clide. Not the best, but pretty decent. He is on TSL which has a pretty shallow player pool so he should see decent play. If he were instead on oGs or Slayers, which have huge, talented Terran pools, he would almost never play. So you have to also consider if the player is unique or good relative to the other players on his team and whether that would allow him to play often. If Clide were on F.United, you would imagine he would play every single match.
ROLES
Another important thing to consider about players is their role on the team. There are a large number of terms used to describe the different functions players can serve on their team, so I'll go ahead and list them here:
Starter
Definition: The guy who will be playing first in a match, usually someone who is among the top 3-4 players on a team and ideally has no ridiculous weakness
Predictability: Probably the hardest to predict who starters will be, at least right now as there is little precedent for teams to have any designated starters but some players do seem to start more than others
Potential Playtime: Obviously they are guaranteed to win
FGSTL Usefulness: Out of all the roles, starters are probably the most useful or at least the most reliable since they will always be able to have an opportunity to win and you don't have to worry about how the match will progress to allow use of the player but only on them winning the first game. If you can somehow determine a player that will always start and is very good at that lead off game, that's basically a guaranteed 2 points each match. However, you have to be careful to choose a good starter because other wise he will drop in trade points and potentially scare his team from using him again (a la Noblesse).
Sniper
Definition: Someone who is trained to play only one specific person, usually the best player on the other team. Reserved for only the very best players such as MC, MVP and such.
Predictability: Unless you have insider info or a person has been publicly announced as a sniper, it's pretty much impossible to predict who this is. In general, this is a team's lower top class player who isn't in GSL and thus hasn't refined their playstyle as much.
Potential Playtime: In order for a sniper to play, the team must play against a team with a notable ace, and force that ace to actually play, meaning the sniper's team must have a better supporting cast than the ace's team. In other words, the thought of a ZENEX sniper for Nestea is pretty laughable as the match probably won't even come close to reaching Nestea
FGSTL Usefulness: Although most people on TL seem to think snipers are among the most important players to pick on your fgstl, I would caution you to at least think about it first. While snipers can rack up huge points for ending streaks and most likely contributing to a team win and thus additional team-player points, consider the fact that the scenario that actually allows for a sniper to play is fairly situational. Also, if the sniper's team loses easily, then he will never see play since the other team's ace won't show up. Also, the sniper may be sent out as soon as the ace wins a game, meaning there is no streak breaking points. Plus, if you pick a sniper for a specific match, what do they do the rest of the season?
Ace
Definition: The best player on the team, typically comes out as their fourth player to salvage a win
Predictability: In almost every case, everyone knows who the ace is. For teams like ZENEX and Prime, who exactly is the ace may be a little ambiguous. New teams like FXO, and FUnited haven't shown who their ace is so they may not have one yet. Don't guess who the ace is. If you have to think about it, you don't know who it is.
Potential Playtime: Since an ace will only play as the last player (generally), you want to pick an ace that will see play time, and ideally, with as few wins as possible by the time it gets to him so that there is a greater potential for consecutive wins. In other words, Nestea is a pretty bad choice, whereas ScfOu would be a pretty solid one.
FGSTL Usefulness: This is probably going to be the biggest risk on your team as you are investing a lot of points on a player that can do really well, but potentially either won't play or won't win when he does due to snipers or other factors. I don't really have enough experience to say that a fteam with no ace is the best, or that a fteam needs an ace, so I'll leave it at that. Just refer back to the first section at the very top when thinking about this.
Specialist
Definition: Compared to a sniper, who practices to take out a specific player, a specialist is someone who is just really good against a specific race.
Predictability: In general you should be able to tell who these are based on TLPD records but you have to be careful to check that the player actually won against good players of that race
Potential Playtime: Probably will see more playtime than a sniper as they are far more versatile in every match and can face any player of that race rather than just that ace on the other team. Some specialists are used as snipers though, which might limit their play in those matches
FGSTL Usefulness: I think they are a good choice since if they are good enough, they will almost always see play. For example, if a terran with bad to only okay TvP is up against Slayers, Alicia is a good bet for seeing play. Since they'll always be sent up against their specialty, whenever they get played they should get at least 1 win.
Meat Shield
Definition: Players that get sent against an ace when there are still games to give in order to tire them out
Predictability: Pretty much anyone who is midtier or lower and wouldn't see play otherwise
Potential Playtime: Depends on the match.
FGSTL Usefulness: Pretty much useless. I really didn't even need to include this role here.
Coach/Caster
Definition: A playable person on the team that has additional roles as the team coach or caster
Predictability: If you don't know who they are for each team, do more research
Potential Playtime: Almost none. Choya is a bit of an exception as he seems to like playing himself relatively often, but don't assume that they will play.
FGSTL Usefulness: This is basically a point trap for newbies that don't study the players very well and waste the slot and 1 point on them.
Potential Stars
Definition: Not really a role per se, but any unknown player that is actually really good but just hasn't shown results yet
Predicatability: Surprisingly, a lot of the "unknowns" from GSTL were actually hyped beforehand, but it's hard to know what is real and who will fluke.
Potential Playtime: They will probably see at least one game early on and if they do well, they will continue to get played.
FGSTL Usefulness: This is probably the most important role on the fteam as this is what will bring in the big points. To have a 1 cost player and get a 10 point return is huge as it allows the rest of your fteam's players to be even better with higher costs. However, this is also a risk as it wastes a precious spot on your fteam. I think that the potential stars are what will make or break the winning teams for this season, as there are a LOT of unknown 1 cost players that have the potential to have a huge season.
TEAMS
The other thing you have to think about picking is the pro team for your fteam. While this follows the same basic principles of player picking of finding where TL made a mistake in the point distribution, there is one huge difference. If a player gets an all-kill, he gains 10 points (2*4+1+1). Meanwhile the team gets 8 points for the 4-0 win. If the player goes 3-0 (he comes in after the starter wins and then loses), he gains 8 points, while the team gets 6 points. So right now, it seems like the team and players going on sweeps seem to scale fairly well.
If you take a look at the scoring diagram for teams, you notice that it is not completely linear:
+ Show Spoiler +
4-0 Victory: 8 Points
4-1 Victory: 6 Points
4-2 Victory: 4 Points
4-3 Victory: 2 Points
3-4 Defeat: 1 Point
2-4 Defeat: 0 Points
1-4 Defeat: -1 Points
0-4 Defeat: -2 Points
4-1 Victory: 6 Points
4-2 Victory: 4 Points
4-3 Victory: 2 Points
3-4 Defeat: 1 Point
2-4 Defeat: 0 Points
1-4 Defeat: -1 Points
0-4 Defeat: -2 Points
At first, the points increase by 1's and then later by 2's in the better half of the results. This means that a team that is twice as good as another team will potentially recieve more than twice the amount of points. The negative points might scare you, but imagine if every scenario received a 2 point boost. This would affect all owners the same (independent of trading), so there's actually no effect of the negative score. The point distribution could look like this:
+ Show Spoiler +
4-0 Victory: 10000008 Points
4-1 Victory: 10000006 Points
4-2 Victory: 10000004 Points
4-3 Victory: 10000002 Points
3-4 Defeat: 10000001 Point
2-4 Defeat: 10000000 Points
1-4 Defeat: 999999 Points
0-4 Defeat: 9999998 Points
4-1 Victory: 10000006 Points
4-2 Victory: 10000004 Points
4-3 Victory: 10000002 Points
3-4 Defeat: 10000001 Point
2-4 Defeat: 10000000 Points
1-4 Defeat: 999999 Points
0-4 Defeat: 9999998 Points
Thus, while the negative score might scare you to think differently about the team, don't. The difference in scaling however, if a different issue. If we were assign 1 costper increase in victories and look at the point/cost ratio, we would see this:
+ Show Spoiler +
4-0 Victory: 8 Points / 8 Cost = 1
4-1 Victory: (6 Points/ 7 Cost) = 0.86
4-2 Victory: (4 Points/ 6 Cost) = 0.67
4-3 Victory: (2 Points/ 5 Cost) = 0.4
3-4 Defeat: (1 Point/ 4 Cost) = 0.25
2-4 Defeat: (0 Points/ 3 Cost) = 0
1-4 Defeat: (-1 Points/ 2 Cost) = -0.5
0-4 Defeat: (-2 Points/ 1 Cost) = -2
4-1 Victory: (6 Points/ 7 Cost) = 0.86
4-2 Victory: (4 Points/ 6 Cost) = 0.67
4-3 Victory: (2 Points/ 5 Cost) = 0.4
3-4 Defeat: (1 Point/ 4 Cost) = 0.25
2-4 Defeat: (0 Points/ 3 Cost) = 0
1-4 Defeat: (-1 Points/ 2 Cost) = -0.5
0-4 Defeat: (-2 Points/ 1 Cost) = -2
As we can see, the better a team is, not only does it bring in more points, but it also is more efficient than worse team.
However, lets assume TL is actually smarter than this an takes the uneven scaling into effect. Lets also add 2 points to the results to get rid of negatives:
+ Show Spoiler +
4-0 Victory: (10 Points / 11 Cost = 0.91
4-1 Victory: (8 Points/ 9 Cost) = 0.89
4-2 Victory: (6 Points/ 7 Cost) = 0.86
4-3 Victory: (4 Points/ 5 Cost) = 0.8
3-4 Defeat: (3 Point/ 4 Cost) = 0.75
2-4 Defeat: (2 Points/ 3 Cost) = 0.66
1-4 Defeat: (1 Points/ 2 Cost) = 0.5
0-4 Defeat: (0 Points/ 1 Cost) = 0
4-1 Victory: (8 Points/ 9 Cost) = 0.89
4-2 Victory: (6 Points/ 7 Cost) = 0.86
4-3 Victory: (4 Points/ 5 Cost) = 0.8
3-4 Defeat: (3 Point/ 4 Cost) = 0.75
2-4 Defeat: (2 Points/ 3 Cost) = 0.66
1-4 Defeat: (1 Points/ 2 Cost) = 0.5
0-4 Defeat: (0 Points/ 1 Cost) = 0
Again, the same general trend. However, this is independent again of the most important part of picking your fteam: you also have to pick players. The main question we now must answer is:
Is it more important to choose a good team or good players?
Looking at last night's results, the players that played and won had these point ratios:
+ Show Spoiler +
DRG: (8 points/8 cost)=1
Alicia: (4 points/5 cost) = 0.8
GuineaPig: (4 points/5 cost) = 0.8
Boxer: (2 points/2 cost) = 1
Alicia: (4 points/5 cost) = 0.8
GuineaPig: (4 points/5 cost) = 0.8
Boxer: (2 points/2 cost) = 1
As it turns out, TL is actually incredible so far when it comes to assigning cost values to players, as they are all nearly 1 or actually perfect. NOT.
Sure, those four players had nearly the exact same point/cost ratio to each other. But what about the other 6 or so 2+ point players on MVP or every other player on Slayers? If you were to actually average out the point/cost ratio of every player on the two teams, it would be a ridiculously low number.
Well, what if you only counted players that would reasonably see play?
+ Show Spoiler +
DRG:8/8=1
GuineaPig:4/5=0.8
Genius:1/4=0.25
HwangSin:1/2=0.5
Keen:1/3-0.33
Noblesse:1/2=0.5
Tails:1/2=0.5
ViOlet:1/2-0.5
Alicia:4/5=0.8
Boxer:2/2=1
Frozen:0
GanZi:0
Golden:0
Min:0
MMA:0
Ryung:0
Taeja:0
YuGiOh:0
Average = 0.34
GuineaPig:4/5=0.8
Genius:1/4=0.25
HwangSin:1/2=0.5
Keen:1/3-0.33
Noblesse:1/2=0.5
Tails:1/2=0.5
ViOlet:1/2-0.5
Alicia:4/5=0.8
Boxer:2/2=1
Frozen:0
GanZi:0
Golden:0
Min:0
MMA:0
Ryung:0
Taeja:0
YuGiOh:0
Average = 0.34
Even if you trim the teams even more and leave only players that definitely will see play:
+ Show Spoiler +
DRG:8/8=1
GuineaPig:4/5=0.8
Genius:1/4=0.25
Keen:1/3-0.33
Noblesse:1/2=0.5
ViOlet:1/2-0.5
Alicia:4/5=0.8
Boxer:2/2=1
GanZi:0
Min:0
MMA:0
Ryung:0
Average = 0.43
GuineaPig:4/5=0.8
Genius:1/4=0.25
Keen:1/3-0.33
Noblesse:1/2=0.5
ViOlet:1/2-0.5
Alicia:4/5=0.8
Boxer:2/2=1
GanZi:0
Min:0
MMA:0
Ryung:0
Average = 0.43
What I conclude from this is, assuming that TL is actually perfect in terms of setting the cost of the teams, the best teams are equally as efficient as the best players and much more efficient than the average results of mid-high range players. You also have to hope that players perform exactly as well each match, and it's entirely possible DRG won't even play in some match, which would be a huge blow in terms of his efficiency compared to teams. Of course, this is assuming that TL is correct in their setting of team costs, which I suspect they are .
IM is rated at 8 cost, meaning they are suspected to go somewhere between 4-1 and 4-2 every game. Though they have a tough matchup with ST, this is not a mind blowing assumption to make for IM's results. Likewise, Slayers is 6 cost, meaning they are suspected to have an average 4-2 to 4-3 victory. While they lost last night, they can easily make up for the slack with some good games in the next few weeks. MVP is 4 cost, meaning they will end up with an average of a 3-4 defeat. If teams can learn to handle DRG somehow with snipers, MVP may end up doing even worse than that.
Therefore, when it comes to teams, while I believe that TL is correct in their cost according to win/loss margins, I believe that they are wrong in their point distribution. In fact, it doesn't even matter if the costs reflect the actual win/loss margins, because ultimately what is important is the number of points you gain. I believe that teams are, at the highest levels, at the same level of efficiency of players.
I also think that teams are far more reliable because you know a team will always be present to play its matches, while a particular player may not always play each week. DRG may not have to play one week, which would totally swing the efficiency towards teams. Another point to note is that while players may slump, teams are usually deep enough to never slump.
ANTI-TEAM
Anti-teams are similar to the main team in that you are looking for the biggest discrepancy of cost to points. However, there is an attitude among some TL'ers that I want to address.
Most people chose players they thought would lose often. This is totally wrong because at worst, a player can score 0 points. While a guy that comes up and loses scores the same as a guy who sat on the bench, the critical difference is that the person playing and losing still has the opportunity to win and score points.
Ideally, you want a player that will never player. Player-team points also indicate that you would prefer him to be on a team that never wins. Of course, this means that they won't be worth much, so you need to balance and compromise those three categories. I think most people chose similar anti-teams so I won't go into much detail about that.
1. Doesn't play (or is bad)
2. On a team that never/rarely wins (FUnited, TSL, FXO)
3. Worth enough cost to fill up minimum.
I feel like most of the actually good teams most of the same basic players for their anti-team:
(Fruitdealer, Trickster, Lyn, GuineaPig, Moon, Maka, Leenock, Killer)
I disagree with Leenock and GuineaPig, if only because they play often. Leenock has done terrible in GSTL before, but the fact of the matter is that he plays all the time for fOu. GuineaPig is just straightup a bad choice and I think people just didn't research him enough to see he wins often. Moon and Lyn are risky because they will play all the time for FUnited, but because they play so late in the season, they may be good for trading later.
That also brings up the issue of potentially trading away anti-team picks if they are up against a weak team where someone like FruitDealer could actually get some wins. As much as he is slumping, I think few would doubt he could put some big damage down on FUnited if it turns out foreigners can't cut it.
I also don't know how big anti-team negative scores get, so we'll see how big a factor player-team scores will be. Since I'm new to this, I don't know how big the margins are for winning teams, so it might be risky to try and cut corners here since the potential for losses are huge.
lol the computer lab I am typing from is closing right now, so I will post this now and get back to it maybe tomorrow or tonight if I have time. I have the other categories I plan to talk about below, and some discussion about my team and why I chose who I did. It's a little sad, but now that I've written all of that, there are some players on my team I may not have chosen now.
DURING THE SEASON
RANDOMNESS
TRADING
point drift
captain
MY TEAM
