|
On October 18 2017 05:01 ppp87 wrote:Show nested quote +On October 18 2017 04:04 conTAgi0n wrote:Damnit Soulkey, what made you think that was a good time to attack with your mutalisks there in that last game? You had the slower spire, faster second hatch, and you knew hero skipped ling speed, so why pick a fight over his natural so early? And after playing so smart in the previous game too... Oh well.. in any case I do think Larva looks better in ZvZ than both hero and Soulkey right now, so he is my pick to make it to the finals. I would have rather watched him face off against Soulkey next round for sure, but c'est la vie. Seriously, that's an incomprehensible move from Soulkey, I just dont understand how it's possible to do it at pro lvl... can someone explain this move please. Nerves
On October 18 2017 11:35 EsportsJohn wrote:Show nested quote +On October 18 2017 05:01 ppp87 wrote:On October 18 2017 04:04 conTAgi0n wrote:Damnit Soulkey, what made you think that was a good time to attack with your mutalisks there in that last game? You had the slower spire, faster second hatch, and you knew hero skipped ling speed, so why pick a fight over his natural so early? And after playing so smart in the previous game too... Oh well.. in any case I do think Larva looks better in ZvZ than both hero and Soulkey right now, so he is my pick to make it to the finals. I would have rather watched him face off against Soulkey next round for sure, but c'est la vie. Seriously, that's an incomprehensible move from Soulkey, I just dont understand how it's possible to do it at pro lvl... can someone explain this move please. Honestly, Soulkey had a lot less information that we did. If we put ourselves in his shoes, it looks like a mirrored build (9pool vs 9pool) in which Soulkey managed to gain map control. Off the back of that map control, he got an earlier natural and what he thought might have been an extra drone or two. With the assumption that both builds were mirrored, there was definitely a chance to bust his opponent there and then just keep flooding scourge until his mutas won out, even with that extra 100 gas that hero saved by not going ling speed (which could have also gone into a fast carapace upgrade and given hero a stronger mid game advantage). The same sort of thing happened in G1 as well. I'm tending toward believing that Soulkey had no idea he was facing an overpool in both games. From his perspective, it would have looked like a mirrored build where he had the tech advantage and a clear timing window to attack. He had seen no speed for hero, then snuck a ling into hero main seeing the faster spire, i doubt he think hero will be having less muta than him, it was just a tilt move, similar to that throw in game 1.
On October 18 2017 05:36 L_Master wrote:Show nested quote +On October 17 2017 23:54 _Animus_ wrote: The best zerg just suicided himself, Flash must be really happy because he just became the new champion. Whats wrong with these youngsters really? Rain and Soulkey having such good play turn into suicide fest. Since when having an eco advantage means u have to desperately do damage to opponent who is getting more and more behind just as the timer is ticking anyway? He got way faster natural and gas in game five, he scouted with ling and knew perfectly that hero have the muta advantage and he just throw his place in the semis with easy hand. Now we have same old guys in the semi finals. If Finals are Hero vs Flash it will be one ugly stomp for Hero. I didnt see Hero being better than soulkey today, he had no better control and notable lack of multitasking compared to soulkey. Soulkey just throw just as Rain throw everything. After my big hope for dethroning flash is gone, dont think there is anything interesting to see after Bisu is out because Flash will toyplay with both Hero/Larva. We already talked about this in the other thread too Animus, Rain didn't throw anything, he just got outplayed. Acting like Flash is more likely to "toyplay" with Larva than he is with SK is absolutely ridiculous. Neither player has a good chance, Flash is Flash when it comes down to it, but Larva, SK, and EffOrt are head and shoulders the best ZvTers playing right now. Moreover acting like Larva's ZvT is even remotely related to herO's ZvT is ridiculous. herO is, and has always been, a decent ZvZer and at times a brilliant ZvPer. His ZvT has always been mediocre. We seem to understand the term outplayed in different ways. For me its just Rain outplayed himself into favor of Larva. Having same number of bases against protoss and droping to almost half of his limit is something that i cannot agree a better player will do against his opponent.
I hope to see Larva doing at least on the same level as Soulkey vs Flash in this tournament, until that happen il have my doubts about him.
|
|
At the rate things are going maybe it makes sense to predict outcomes using 6-month spon match statistics (5/17 to 10/17):
Ro8
Larva vs Rain: 31 Games, 20W 11L = 64.5%
Larva dominating victory (fulfilled)
Hero vs Soulkey: 17 Games, 9W 8L = 52.9%
Hero slight edge victory (fulfilled)
Bisu vs killer: 16 Games, 14W 2L = 87.5%
Bisu dominating victory - prediction
Flash vs Mind: 14 Games, 11W 3L = 78.6%
Flash dominating victory - prediction
---
Ro4
hero vs Larva: 7 Games, 6W 1L = 85.7%
Hero dominating victory - prediction
Flash vs Bisu: 27 Games, 18W 9L = 66.7%
Flash dominating victory - prediction
---
Finals
Flash vs hero: 60 Games, 49W 11L = 81.7%
Flash dominating victory - prediction
|
On October 18 2017 19:19 orvinreyes wrote: At the rate things are going maybe it makes sense to predict outcomes using 6-month spon match statistics (5/17 to 10/17):
No, it doesn't work like that, we can not reliably predict future events based on data from the past, the future doesn't work like that. This is like trying to predict how severe the next financial crisis will be. You just can't. You stress test the banks with crisis from the past but no, it still sucks for prediction. Because the future one will somehow always be stronger. It's the same in sports. This team wins 80% of games past so it will win in the finals. No, no, no. The team might win but it's not because of past victories, this is not a reliable point on which to base future predictions. Never has been, never will be.
|
Russian Federation19 Posts
On October 18 2017 19:40 IntoTheStorm wrote:Show nested quote +On October 18 2017 19:19 orvinreyes wrote: At the rate things are going maybe it makes sense to predict outcomes using 6-month spon match statistics (5/17 to 10/17):
No, it doesn't work like that, we can not reliably predict future events based on data from the past, the future doesn't work like that. This is like trying to predict how severe the next financial crisis will be. You just can't. You stress test the banks with crisis from the past but no, it still sucks for prediction. Because the future one will somehow always be stronger. It's the same in sports. This team wins 80% of games past so it will win in the finals. No, no, no. The team might win but it's not because of past victories, this is not a reliable point on which to base future predictions. Never has been, never will be.
Well, we can, to some degree, but the prediction does not work for a single match or even for a bo 5, as you correctly point out. We can estimate, given a large sample of games that, say, Bisu's true skill in PvZ in X%, and Flash's true skill in TvZ is Y%, but it doesn't of course mean that they win their next BO3 or BO5 3-1 or 3-2, because it's just too much variation. It means that in the long run they are most likely to have X% or Y% winrate in a particular matchup.
Needless to say, the data sample cited above is way too small. Besides, ZvZ is probably much much harder to predict and needs probably x3 or x4 size of a sample to atleast start talking about estimating the players.
Most probably, the data we have is sufficient only for determining how one particular player does in a certain matchup. We will most probably never be able to reliably evaluate player vs. player data, although if Larva sometime in the future amasses 1000 recorded games vs. Flash and 1000 vs. Bisu, maybe it will be possible then.
|
On October 18 2017 19:49 Skybrod wrote: although if Larva sometime in the future amasses 1000 recorded games vs. Flash and 1000 vs. Bisu, maybe it will be possible then.
No, it would still not be possible, regardless of the sample size, 1000 or 10000 games vs a said player. You would only know that Larva has won like 40% of the past games vs Bisu.Or 50%, 60%, whatever. You can not predict the outcome of the next game based on that. It should be clear to everyone. There is randomness, there are other things that affect how a player plays a game, there is health, there is attitude at the time of the game, there is mentality - too many factors to reliably predict anything. And this is the beauty of all competition - it is unpredictable, rare events do happen and influence the outcome of competition. This unpredictability is what bookmakers make their money off of. For spectators like us, this quality of the games is amazing - to not know when the favorite will slip, this is what makes every sport (electronic of physical) attractive to me. This is why I watch muay thai matches - there is this dude, he is probably better but does it mean an auto win? No, the other guy is there to challenge and test him. He might connect with an elbow or a head kick, you never know.
|
On October 18 2017 20:12 IntoTheStorm wrote:Show nested quote +On October 18 2017 19:49 Skybrod wrote: although if Larva sometime in the future amasses 1000 recorded games vs. Flash and 1000 vs. Bisu, maybe it will be possible then. No, it would still not be possible, regardless of the sample size, 1000 or 10000 games vs a said player. You would only know that Larva has won like 40% of the past games vs Bisu.Or 50%, 60%, whatever. You can not predict the outcome of the next game based on that. It should be clear to everyone. There is randomness, there are other things that affect how a player plays a game, there is health, there is attitude at the time of the game, there is mentality - too many factors to reliably predict anything. And this is the beauty of all competition - it is unpredictable, rare events do happen and influence the outcome of competition. This unpredictability is what bookmakers make their money off of. For spectators like us, this quality of the games is amazing - to not know when the favorite will slip, this is what makes every sport (electronic of physical) attractive to me. This is why I watch muay thai matches - there is this dude, he is probably better but does it mean an auto win? No, the other guy is there to challenge and test him. He might connect with an elbow or a head kick, you never know.
Fully agree, correlation != causation. I played enough poker and bet on enough stocks and fx trades to know. Those are just my predictions for fun. Sorry if I shit on the sweet science of statistics.
|
Croatia9362 Posts
On October 18 2017 19:40 IntoTheStorm wrote:Show nested quote +On October 18 2017 19:19 orvinreyes wrote: At the rate things are going maybe it makes sense to predict outcomes using 6-month spon match statistics (5/17 to 10/17):
No, it doesn't work like that, we can not reliably predict future events based on data from the past, the future doesn't work like that. You've just described whole of science. Saying it doesn't work is quite funny.
|
Russian Federation19 Posts
On October 18 2017 20:12 IntoTheStorm wrote:Show nested quote +On October 18 2017 19:49 Skybrod wrote: although if Larva sometime in the future amasses 1000 recorded games vs. Flash and 1000 vs. Bisu, maybe it will be possible then. No, it would still not be possible, regardless of the sample size, 1000 or 10000 games vs a said player. You would only know that Larva has won like 40% of the past games vs Bisu.Or 50%, 60%, whatever. You can not predict the outcome of the next game based on that. It should be clear to everyone. There is randomness, there are other things that affect how a player plays a game, there is health, there is attitude at the time of the game, there is mentality - too many factors to reliably predict anything. And this is the beauty of all competition - it is unpredictable, rare events do happen and influence the outcome of competition. This unpredictability is what bookmakers make their money off of. For spectators like us, this quality of the games is amazing - to not know when the favorite will slip, this is what makes every sport (electronic of physical) attractive to me. This is why I watch muay thai matches - there is this dude, he is probably better but does it mean an auto win? No, the other guy is there to challenge and test him. He might connect with an elbow or a head kick, you never know.
I did not claim that one could predict the outcome of the next game, I explicitly stated that it is not possible, I was talking about the estimate of a player's perfomance in the long run, given sufficiently large sample size. That's how projections in baseball nowadays work. They project a player's whole season (400-500 plate appearances), although baseball is probably better suited for analysis than BW matches.
I would also disagree with you on your claim that prediction is not possible based on previous data. There's such thing as statistical learning which is applied to many spheres. E.g. if the size of tumor strongly correlates with the fact whether this tumor is malign or benign, we can make a sufficiently accurate prediction that if a person has a tumor of certain size, there's a good chance it will be malign/benign. Sorry for the offtop, just to illustrate the point.
|
Good job Hero! I like Soulkey a lot and expect him to do very well in the future, but today Hero played better. Moreover, he played better in a tricky way, which is exactly what a champion must do.
His overlord-dodging on Crossing Field was awesome.
|
On October 18 2017 20:45 2Pacalypse- wrote:Show nested quote +On October 18 2017 19:40 IntoTheStorm wrote:On October 18 2017 19:19 orvinreyes wrote: At the rate things are going maybe it makes sense to predict outcomes using 6-month spon match statistics (5/17 to 10/17):
No, it doesn't work like that, we can not reliably predict future events based on data from the past, the future doesn't work like that. You've just described whole of science. Saying it doesn't work is quite funny.
User was warned for this post
|
United States4883 Posts
On October 18 2017 19:40 IntoTheStorm wrote:Show nested quote +On October 18 2017 19:19 orvinreyes wrote: At the rate things are going maybe it makes sense to predict outcomes using 6-month spon match statistics (5/17 to 10/17):
No, it doesn't work like that, we can not reliably predict future events based on data from the past, the future doesn't work like that.5
Lolwut
|
Russian Federation483 Posts
On October 18 2017 07:14 FlaShFTW wrote: To everyone who's constantly moaning and complaining about a player upsetting another player because he wont have a chance against FlaSh in the finals: shut up. There is no "should have". There is only "happens". A player lost. Maybe he didn't have his A game that day. Maybe the other player was playing stellar that day. Who knows? But to comment that a player shouldn't have won because he's going to get owned further down the tourney, then why play at all? Why would hero even play this tournament knowing he has no chance against FlaSh? Why would 4/5ths of this field play if they knew they had no chance against FlaSh in a BOX down the line? That attitude is terrible, and it undermines the level of practice and skill that the players have had to develop to participate in this tournament. Shame on anyone who tries to belittle or put down players simply because they can't beat FlaSh in BoX. amen
|
I thought Soul key actually player really well, the 2-3 result is actually remarkable that every game he either had a scouting disadvantage with mirror builds or a build order disadvantage.
|
On October 18 2017 23:05 EsportsJohn wrote:Show nested quote +On October 18 2017 19:40 IntoTheStorm wrote:On October 18 2017 19:19 orvinreyes wrote: At the rate things are going maybe it makes sense to predict outcomes using 6-month spon match statistics (5/17 to 10/17):
No, it doesn't work like that, we can not reliably predict future events based on data from the past, the future doesn't work like that.5 Lolwut
I'm not sure what he is trying to say either. If he is just saying that you can predict, as in know for certain the outcome, then yea that's certainly true but also incredibly obvious. I think that's what he is trying to say because in his quote below he says "this team wins 80% of games past so it WILL win the finals".
Obviously, we can't use the word will, and I think that's all he is trying to say.
If he is trying to say that you can't use data to predict the likelihood of a particular outcome...then he is just off the rails crazy.
On October 18 2017 19:40 IntoTheStorm wrote: This team wins 80% of games past so it will win in the finals. No, no, no. The team might win but it's not because of past victories, this is not a reliable point on which to base future predictions. Never has been, never will be.
No they won't win because of past victories. + Show Spoiler +Well, at least considering what we are talking about. Obviously past victories can improve a player by increasing confidence, offering opportunites to improve strategy, etc. but that's not what this discussion is about. In fairness, that's also incredibly obvious.
On the other hand we absolutely can use past results to predict the chances of a particular outcome. If Flash beat's Bisu 65% of the time, you're right it doesn't mean Flash WILL beat Bisu this time in a single game, a Bo5, or even a Bo99. What it does mean is that Flash's odds of winning against Bisu in a Bo5 are about 77%.
Interestingly, if you assume Flash's WR is an even 70% his odds of winning from Ro8 on out are about 60%. His chances of winning the tournament from the Ro16 are a little trickier because group stage is not a Bo5, but would likely be around 45%.
|
Bulgaria750 Posts
On October 18 2017 21:36 Skybrod wrote:Show nested quote +On October 18 2017 20:12 IntoTheStorm wrote:On October 18 2017 19:49 Skybrod wrote: although if Larva sometime in the future amasses 1000 recorded games vs. Flash and 1000 vs. Bisu, maybe it will be possible then. No, it would still not be possible, regardless of the sample size, 1000 or 10000 games vs a said player. You would only know that Larva has won like 40% of the past games vs Bisu.Or 50%, 60%, whatever. You can not predict the outcome of the next game based on that. It should be clear to everyone. There is randomness, there are other things that affect how a player plays a game, there is health, there is attitude at the time of the game, there is mentality - too many factors to reliably predict anything. And this is the beauty of all competition - it is unpredictable, rare events do happen and influence the outcome of competition. This unpredictability is what bookmakers make their money off of. For spectators like us, this quality of the games is amazing - to not know when the favorite will slip, this is what makes every sport (electronic of physical) attractive to me. This is why I watch muay thai matches - there is this dude, he is probably better but does it mean an auto win? No, the other guy is there to challenge and test him. He might connect with an elbow or a head kick, you never know. I did not claim that one could predict the outcome of the next game, I explicitly stated that it is not possible, I was talking about the estimate of a player's perfomance in the long run, given sufficiently large sample size. That's how projections in baseball nowadays work. They project a player's whole season (400-500 plate appearances), although baseball is probably better suited for analysis than BW matches. I would also disagree with you on your claim that prediction is not possible based on previous data. There's such thing as statistical learning which is applied to many spheres. E.g. if the size of tumor strongly correlates with the fact whether this tumor is malign or benign, we can make a sufficiently accurate prediction that if a person has a tumor of certain size, there's a good chance it will be malign/benign. Sorry for the offtop, just to illustrate the point.
This discussion is way too philosophical, to be honest. Not a bad thing, but all points are valid logically and impossible to prove, since at a certain level EVERYTHING is unpredictable, including my own involvement with this thread. A roof tile may fall on my head when I leave home. I really hope not, though.
EDIT: More seriously, though, statistics are of no practical use in this scenario. No progamer would go into the match unprepared because he has the better record against the opponent; I seriously doubt intimidation or complacency based on records is a factor for experienced players either. Besides betting with money on the line, there's no use for statistics for us fans. We can argue all we want and I for example will still have my chest pounding in every game Flash is playing from now on.
In my job we use statistical models all the time, mostly for risk stratification. And, of course, no one ever goes to work resigned to a negative outcome just because his last 100 patients survived, so this one won't.
|
great series! zvz looks reeeally fricken hard O)_O
|
On October 18 2017 20:12 IntoTheStorm wrote:Show nested quote +On October 18 2017 19:49 Skybrod wrote: although if Larva sometime in the future amasses 1000 recorded games vs. Flash and 1000 vs. Bisu, maybe it will be possible then. No, it would still not be possible, regardless of the sample size, 1000 or 10000 games vs a said player. You would only know that Larva has won like 40% of the past games vs Bisu.Or 50%, 60%, whatever. You can not predict the outcome of the next game based on that. It should be clear to everyone.
It is clear to everyone...that you can make predictions about what will happen. In fact, making a prediction is simply saying "I think Bisu will win". I can do that with no evidence at all. Pretty hollow predication, but I could make it.
Again, if you mean know for certain who will win, obviously not. But no one in there right mind would ever think you could know for certain who was going to win based on past results.
What you CAN do, is state the probability of certain results. As outlined before, if Flash wins 80% of his games versus hero, his odds of winning a Bo5 against him are about 94%. This is where a large sample size can help increase confidence in those odds. If Flash had just appeared and won 4-1 against Bisu we wouldn't know if he just had the most amazing day of his life or is that was fairly typical play. But if Flash has played against Bisu for 100k games, then we have a reasonably sampling of conditions taking into account good days/bad days that tells us in general how Flash performs against Bisu. It says if both players are on an average day, this is the likelihood of winning.
As you have mentioned, there are good days and bad days; aka variation. If Flash is sick, he might lose; which we know is a possible outcome because Flash has a 23% chance of losing a Bo5 to Bisu if he wins 65% of their games.
You even know this intrinsically as you stated:
This is why I watch muay thai matches - there is this dude, he is probably better but does it mean an auto win?
You understand this concept intrinsically. t's no different that saying that FlaSh is a better starcraft player than Artosis. If I asked you think is going to win if Flash played Artosis, you would obviously answer FlaSh; same as you would if I offered you $100k if you correctly guess the winner of a Bo5 between FlaSh and Artosis.
The fact that you'd pick FlaSh means you think he is more likely to win. Well...why do you think that? Certainly you don't know anything about how ready Flash is to play right now or how his condition would be in two days, so you can't be making your decision based on that. What's left? Past results, performance, and gameplay. You've just used FlaSh's past starcraft to make a future decision about who is more likely to win and which player you should pick to optimize your chances to earn $100k.
|
Russian Federation19 Posts
On October 19 2017 03:54 L_Master wrote:
On the other hand we absolutely can use past results to predict the chances of a particular outcome. If Flash beat's Bisu 65% of the time, you're right it doesn't mean Flash WILL beat Bisu this time in a single game, a Bo5, or even a Bo99. What it does mean is that Flash's odds of winning against Bisu in a Bo5 are about 77%.
Interestingly, if you assume Flash's WR is an even 70% his odds of winning from Ro8 on out are about 60%. His chances of winning the tournament from the Ro16 are a little trickier because group stage is not a Bo5, but would likely be around 45%.
I guess what he is trying to say is that the outcome of a single game or a short series of games cannot be predicted accurately on the basis of the previous data, and that is true.
|
On October 19 2017 03:55 TaardadAiel wrote:Show nested quote +On October 18 2017 21:36 Skybrod wrote:On October 18 2017 20:12 IntoTheStorm wrote:On October 18 2017 19:49 Skybrod wrote: although if Larva sometime in the future amasses 1000 recorded games vs. Flash and 1000 vs. Bisu, maybe it will be possible then. No, it would still not be possible, regardless of the sample size, 1000 or 10000 games vs a said player. You would only know that Larva has won like 40% of the past games vs Bisu.Or 50%, 60%, whatever. You can not predict the outcome of the next game based on that. It should be clear to everyone. There is randomness, there are other things that affect how a player plays a game, there is health, there is attitude at the time of the game, there is mentality - too many factors to reliably predict anything. And this is the beauty of all competition - it is unpredictable, rare events do happen and influence the outcome of competition. This unpredictability is what bookmakers make their money off of. For spectators like us, this quality of the games is amazing - to not know when the favorite will slip, this is what makes every sport (electronic of physical) attractive to me. This is why I watch muay thai matches - there is this dude, he is probably better but does it mean an auto win? No, the other guy is there to challenge and test him. He might connect with an elbow or a head kick, you never know. I did not claim that one could predict the outcome of the next game, I explicitly stated that it is not possible, I was talking about the estimate of a player's perfomance in the long run, given sufficiently large sample size. That's how projections in baseball nowadays work. They project a player's whole season (400-500 plate appearances), although baseball is probably better suited for analysis than BW matches. I would also disagree with you on your claim that prediction is not possible based on previous data. There's such thing as statistical learning which is applied to many spheres. E.g. if the size of tumor strongly correlates with the fact whether this tumor is malign or benign, we can make a sufficiently accurate prediction that if a person has a tumor of certain size, there's a good chance it will be malign/benign. Sorry for the offtop, just to illustrate the point. This discussion is way too philosophical, to be honest. Not a bad thing, but all points are valid logically and impossible to prove, since at a certain level EVERYTHING is unpredictable, including my own involvement with this thread. A roof tile may fall on my head when I leave home. I really hope not, though.
This is actually a pretty argued topic.
One strong belief is that at a certain level everything is completely predictable and determined. If you know all the initial conditions of the universe of all particles, and apply the laws of physics to them, there is only one way they can behave and thus one possible outcome.
|
|
|
|