That is just patently not true. There are certainly matchups that are 'determined by the die roll,' but that only applies if both players know both decks strengths and weaknesses, as well as the likely contents of an opponent's deck and hand, and both players play optimally.
That said, your point about FNM is well made. Luck certainly has a lot more to do with the outcome when any tournament has a median skill level on the lower side.
Luck in every game of magic:
Your opening 7,
Your opening 6, Your opening 5,
Your first draw,
Your next draw, Every card drawn thereafter.
Getting your outs,
Deciding to bring an out forthis match up.
Playing to the specific out you're going to draw.
Your T1 play in the dark,
Playing around cards,
Playing into cards,
And the list goes on and on.
Yes you can win games with skill, but more often than not games are determined the moment after your opponent cuts your deck.
You might have constructed formats confused with Momir Basic or something.
Modern decks are composed of 60 selected cards, not 60 cards chosen at random. Your opening 7 or 6 or 5 or whatever are determined by how many of each card you are running which affects the probability that you will see them in your opening hand.
And because I'm not playing manaless dredge I need to have some luck on my side in order to draw a hand with both land and spells. Sure I can influence the odds by the way I build my deck, but not hitting the 4% chance that I draw zero lands is luck.
Playing into or around cards aren't luck either. Case in point, my opponent does a lot of draw go stuff and casts and Elesh Norn. On my turn, I fire off a Slaughter Games naming Gifts Ungiven which my opponent didn't show me in game 1. I proceed to thumb through his deck and remove all 4 copies. As I was doing so, he asked me if I saw his deck before we started playing, to which I replied 'You're playing Elesh Norn.' It was not a 'lucky' guess.
So your case and point is that you made an assumption about how a deck was built based on one card and because you happened to be right there must not be an element of luck involved? What if it wasn't using Gifts as the enabler, but some other card? I doubt we'd see you posting about the skill play of getting matched up against a rouge build?
The very next round my T1 play was to cycle two Street Wraiths and play a Blackcleave Cliffs and my opponent goes 'Jund Death's Shadow or Living End?' Again, there was no luck involved in knowing what he was up against. It was experience and repetitions playing many matches. And skill.
He's going to be so mad when you're actually using Streetwraith to cycle through to a fast-mana combo Guild Feud or anything else. Lucky him, I guess?
Playing into or around cards aren't luck either. Case in point, my opponent does a lot of draw go stuff and casts and Elesh Norn. On my turn, I fire off a Slaughter Games naming Gifts Ungiven which my opponent didn't show me in game 1. I proceed to thumb through his deck and remove all 4 copies. As I was doing so, he asked me if I saw his deck before we started playing, to which I replied 'You're playing Elesh Norn.' It was not a 'lucky' guess.
That's cool and all, but the "luck" factor isn't your ability to guess correctly, it's whether or not your opponent actually has what you think he has. I have been Slaughter Games'd for some wildly irrelevant stuff (and total misses) when I was in my URx brewing phase. Based on the cards my opponent had seen, they usually made exactly the right call in what to name, but based on the luck of being matched up with me, instead of a more recognizable decklist, their SG result ranged from weak to whiff. Sometimes, even the "right" call is still wrong, and it's wrong because of luck.
Playing into or around cards aren't luck either. Case in point, my opponent does a lot of draw go stuff and casts and Elesh Norn. On my turn, I fire off a Slaughter Games naming Gifts Ungiven which my opponent didn't show me in game 1. I proceed to thumb through his deck and remove all 4 copies. As I was doing so, he asked me if I saw his deck before we started playing, to which I replied 'You're playing Elesh Norn.' It was not a 'lucky' guess.
That's cool and all, but the "luck" factor isn't your ability to guess correctly, it's whether or not your opponent actually has what you think he has. I have been Slaughter Games'd for some wildly irrelevant stuff (and total misses) when I was in my URx brewing phase. Based on the cards my opponent had seen, they usually made exactly the right call in what to name, but based on the luck of being matched up with me, instead of a more recognizable decklist, their SG result ranged from weak to whiff. Sometimes, even the "right" call is still wrong, and it's wrong because of luck.
I mean, I suppose that's going to be a luck-influenced miss in some very, very tiny % of games. In the vast majority of examples, however, when an opponent guesses Wraiths/Cliffs are most likely to be either Jund DS or Living End, they are mostly going to be correct. Or when they SG Snapcaster or Bolt, that's probably a hit in a huge majority of cases.
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
Playing into or around cards aren't luck either. Case in point, my opponent does a lot of draw go stuff and casts and Elesh Norn. On my turn, I fire off a Slaughter Games naming Gifts Ungiven which my opponent didn't show me in game 1. I proceed to thumb through his deck and remove all 4 copies. As I was doing so, he asked me if I saw his deck before we started playing, to which I replied 'You're playing Elesh Norn.' It was not a 'lucky' guess.
That's cool and all, but the "luck" factor isn't your ability to guess correctly, it's whether or not your opponent actually has what you think he has. I have been Slaughter Games'd for some wildly irrelevant stuff (and total misses) when I was in my URx brewing phase. Based on the cards my opponent had seen, they usually made exactly the right call in what to name, but based on the luck of being matched up with me, instead of a more recognizable decklist, their SG result ranged from weak to whiff. Sometimes, even the "right" call is still wrong, and it's wrong because of luck.
I mean, I suppose that's going to be a luck-influenced miss in some very, very tiny % of games. In the vast majority of examples, however, when an opponent guesses Wraiths/Cliffs are most likely to be either Jund DS or Living End, they are mostly going to be correct. Or when they SG Snapcaster or Bolt, that's probably a hit in a huge majority of cases.
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
Playing into or around cards aren't luck either. Case in point, my opponent does a lot of draw go stuff and casts and Elesh Norn. On my turn, I fire off a Slaughter Games naming Gifts Ungiven which my opponent didn't show me in game 1. I proceed to thumb through his deck and remove all 4 copies. As I was doing so, he asked me if I saw his deck before we started playing, to which I replied 'You're playing Elesh Norn.' It was not a 'lucky' guess.
That's cool and all, but the "luck" factor isn't your ability to guess correctly, it's whether or not your opponent actually has what you think he has. I have been Slaughter Games'd for some wildly irrelevant stuff (and total misses) when I was in my URx brewing phase. Based on the cards my opponent had seen, they usually made exactly the right call in what to name, but based on the luck of being matched up with me, instead of a more recognizable decklist, their SG result ranged from weak to whiff. Sometimes, even the "right" call is still wrong, and it's wrong because of luck.
Your example doesn't make a lot of sense. If your deck is way outside the norm, that isn't unlucky for your opponent, it's a good strategic play by you so long as your brew isn't total garbage that folds to commonly played decks.
All that said, I think your view on luck and Magic are ridiculously overblown. There is no way the data backs you up. The reason Owen Turtenwald won back to back GPs and made 7 top 8s in one year over 4 different formats was not because he was really lucky. It's because he was/is the best player in the world. If you or I played 10 matches against Owen where our decklist was favored 65/35, he would beat us 7 out of 10 times.
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
Well, actually, 50% luck can still work with consistent player performance.
If 50% of the game is luck, that would mean that, logically, the other half is skill. Therefore, with sufficient skill, that remaining 50% would be completely filled, leaving you with 50% to fill that's luck. So you "start out" at 50%, and then would (randomly) get somewhere from 0-50% on the luck portion, averaging out to 25%. Put them together and you get 75%, meaning a median of 11.25 wins at a Grand Prix (15*0.75=11.25). A median of 11 out of 15 wins seems like enough to do well consistently.
And let's not forget certain advantages that the "pros" have, namely guaranteed byes at Grand Prix. Silver gets you 2 guaranteed byes and Gold/Platinum gets you 3. If we use the same 75% as above, but apply it to the 12 games they'd actually play and then add in the 3 guaranteed wins, that gives them a median of 12 wins, certainly enough for considerable consistency in succeeding. (if they have 2 byes, that gives a median of 11.75 wins)
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
Well, actually, 50% luck can still work with consistent player performance.
If 50% of the game is luck, that would mean that, logically, the other half is skill. Therefore, with sufficient skill, that remaining 50% would be completely filled, leaving you with 50% to fill that's luck. So you "start out" at 50%, and then would (randomly) get somewhere from 0-50% on the luck portion, averaging out to 25%. Put them together and you get 75%, meaning a median of 11.25 wins at a Grand Prix (15*0.75=11.25). A median of 11 out of 15 wins seems like enough to do well consistently.
And let's not forget certain advantages that the "pros" have, namely guaranteed byes at Grand Prix. Silver gets you 2 guaranteed byes and Gold/Platinum gets you 3. If we use the same 75% as above, but apply it to the 12 games they'd actually play and then add in the 3 guaranteed wins, that gives them a median of 12 wins, certainly enough for considerable consistency in succeeding. (if they have 2 byes, that gives a median of 11.75 wins)
I deliberately did not use GP as examples because it's hard to sort out the bye issue. My guess is that pros would do just fine without the byes but it's hard to prove. That's why I cited MTGO finishes. There are plenty of regulars whose consistent 5-0s can't be chalked up to luck, especially when it is with certain decks that benefit from luck much less than others.
Luck and skill are very closely related in card games
I think a lot of the skill of mtg is learning how to deal with the luck factor. If you have to mull to 6, you got unlucky, but knowing that you have to mull takes skill. You have plays where you should play around the last possible out for your opponent to turn a 95% win into a 100% win, but sometimes it is better to just pray they don't draw it and go in for the kill.
The only way to resolve this would be performing a statistical analysis using a large sample size, with say 3 catagories good, medium and noob players (enough to know the mechanics) with extremely balanced or exactly same decks. With the results one can run one The relevant statistical procedure to obtain a result. Someone with enough friends please conduct to resolve this question that appears to come up regularly. I am even more than happy to help run the stats. Yes there is some skill and yes there is some luck both are obviously there. What is important really is the level of perceived luck (if real or not) acceptable to a player to keep engaging in playing. Moreover looking at the core principle of mtg which is card collection and deck building is not the fun really in building decks and seeing how they perform
The only way to resolve this would be performing a statistical analysis using a large sample size, with say 3 catagories good, medium and noob players (enough to know the mechanics) with extremely balanced or exactly same decks. With the results one can run one The relevant statistical procedure to obtain a result. Someone with enough friends please conduct to resolve this question that appears to come up regularly. I am even more than happy to help run the stats. Yes there is some skill and yes there is some luck both are obviously there. What is important really is the level of perceived luck (if real or not) acceptable to a player to keep engaging in playing. Moreover looking at the core principle of mtg which is card collection and deck building is not the fun really in building decks and seeing how they perform
That would probably work but would also be unrealistically labor intensive. It would also fail to account for the complexities of real Modern or other format metagames.
All the data one needs to analyze this is on the MTG Elo Project website. Just see how often better players win/lose against worse players in Modern GP and see if that's higher, lower, or at the same rate as expected by their rating. Still labor intensive but it's all math and looking up players, not actually playing thousands of non-representative games.
The only way to resolve this would be performing a statistical analysis using a large sample size, with say 3 catagories good, medium and noob players (enough to know the mechanics) with extremely balanced or exactly same decks. With the results one can run one The relevant statistical procedure to obtain a result. Someone with enough friends please conduct to resolve this question that appears to come up regularly. I am even more than happy to help run the stats. Yes there is some skill and yes there is some luck both are obviously there. What is important really is the level of perceived luck (if real or not) acceptable to a player to keep engaging in playing. Moreover looking at the core principle of mtg which is card collection and deck building is not the fun really in building decks and seeing how they perform
That would probably work but would also be unrealistically labor intensive. It would also fail to account for the complexities of real Modern or other format metagames.
All the data one needs to analyze this is on the MTG Elo Project website. Just see how often better players win/lose against worse players in Modern GP and see if that's higher, lower, or at the same rate as expected by their rating. Still labor intensive but it's all math and looking up players, not actually playing thousands of non-representative games.
Would you mind doing the legwork by any chance? It would be a fantastic article for this site and would help a lot of players better understand the nature of MTG.
I know I'm the OP and I should probably be the one to do it, but I honestly just don't have time with work and everything else.
The long and short of it is that there is some non-zero amount of luck that decides all games of Magic. Some amount of games that, no matter your skill, no matter your deck, no matter your card draw, and no matter your matchup, you WILL lose. These losses can be mitigated by some amount of skill in deck choice, deck construction, and play decisions, but that exact percentage by which your win rate goes up is up to interpretation and very difficult to decipher because of the vast number of variables (not the least bit due to what is considered "correct" or not for any given play, since plays are made without perfect information). Yes, skill plays a role in increasing win percentage (I don't think anyone is disputing this), but what would be interesting is to see by how much. However, that's difficult to do using big events because, as mentioned before, many of the best players start out with free byes that give them free wins and help them dodge many of the strange and weird matchups they may otherwise lose. Without a massive number crunch from events unaffected by byes, pretty much any reasonable conclusion will be based on feelings and anecdotes.
The long and short of it is that there is some non-zero amount of luck that decides all games of Magic. Some amount of games that, no matter your skill, no matter your deck, no matter your card draw, and no matter your matchup, you WILL lose. These losses can be mitigated by some amount of skill in deck choice, deck construction, and play decisions, but that exact percentage by which your win rate goes up is up to interpretation and very difficult to decipher because of the vast number of variables (not the least bit due to what is considered "correct" or not for any given play, since plays are made without perfect information). Yes, skill plays a role in increasing win percentage (I don't think anyone is disputing this), but what would be interesting is to see by how much. However, that's difficult to do using big events because, as mentioned before, many of the best players start out with free byes that give them free wins and help them dodge many of the strange and weird matchups they may otherwise lose. Without a massive number crunch from events unaffected by byes, pretty much any reasonable conclusion will be based on feelings and anecdotes.
Like I said, the answer to this is on the MTG Elo project page. I even explained how to do it above. This will tell you based on real games whether or not players are winning in Modern as their rating predicts, or less/more than their rating predicts. If you have a bunch of players who are supposed to win 64% of their matches against the opponent based on rating, but then they are winning only 55% of matches, that's a pretty strong indicator that luck is at play in that other 9%. N is large enough, it's real world data, and it's literally all on the site. Someone just needs to crunch the numbers. I am not going to do it, but it's a great article/post/thread opportunity for anyone who wants to do the research and the math.
But again, it's not some unknowable element. It's very knowable and calculable. It will just be time consuming.
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
Well, actually, 50% luck can still work with consistent player performance.
If 50% of the game is luck, that would mean that, logically, the other half is skill. Therefore, with sufficient skill, that remaining 50% would be completely filled, leaving you with 50% to fill that's luck. So you "start out" at 50%, and then would (randomly) get somewhere from 0-50% on the luck portion, averaging out to 25%. Put them together and you get 75%, meaning a median of 11.25 wins at a Grand Prix (15*0.75=11.25). A median of 11 out of 15 wins seems like enough to do well consistently.
And let's not forget certain advantages that the "pros" have, namely guaranteed byes at Grand Prix. Silver gets you 2 guaranteed byes and Gold/Platinum gets you 3. If we use the same 75% as above, but apply it to the 12 games they'd actually play and then add in the 3 guaranteed wins, that gives them a median of 12 wins, certainly enough for considerable consistency in succeeding. (if they have 2 byes, that gives a median of 11.75 wins)
I deliberately did not use GP as examples because it's hard to sort out the bye issue. My guess is that pros would do just fine without the byes but it's hard to prove. That's why I cited MTGO finishes. There are plenty of regulars whose consistent 5-0s can't be chalked up to luck, especially when it is with certain decks that benefit from luck much less than others.
For the best pro's having 3 byes is enormous. While the average competitor you'll face in the first 2-3 rounds is much lower than in the later rounds, modern is such a MU-dependent format and the range of decks being played on Day 1's Rounds 1-3 is so much wider than later in the tournament that the odds of you running up against a terrible MU increases significantly (unless of course you have terrible T1 MU's...). I can't stress enough how much of an advantage having 2 byes is, let alone 3! It's much much easier to go 10-2 dodging a wide field in the first 2-3 rounds than it is going 13-2 with no byes. It's also the main reason (imho) why you see the top players be more consistent on the GP circuit. I know very very good players who consistently put up 11-4 and 10-5 records with 0-1 byes, but they don't really grind GP's so don't get the necessary pro points or PWP for 2+ byes. I have a feeling WoTC does this because they know that the best players will hardly come T32 in 15+ round 2000+ player tournaments with no byes. GP's are all ready terribly poor EV, imagine having no byes as well. I doubt you'd see all the pros going to so many GP's without their 2 or 3 byes.
My point is, that a lot of people overplay the amount of luck in Modern, but underplay MU "luck" and the cumulative effect of luck for premier and GP level events that are 15+ rounds (especially with no byes). Or in other words, there is too much variance for even the best players in the world to consistently put up money finishes in GP level events without byes, which tells you a lot.
TL;DR The only way to know for sure whether "pros" are good enough to see consistently at the top tables at GP level events is to get rid of byes. They skew the data WAY too much.
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
Well, actually, 50% luck can still work with consistent player performance.
If 50% of the game is luck, that would mean that, logically, the other half is skill. Therefore, with sufficient skill, that remaining 50% would be completely filled, leaving you with 50% to fill that's luck. So you "start out" at 50%, and then would (randomly) get somewhere from 0-50% on the luck portion, averaging out to 25%. Put them together and you get 75%, meaning a median of 11.25 wins at a Grand Prix (15*0.75=11.25). A median of 11 out of 15 wins seems like enough to do well consistently.
And let's not forget certain advantages that the "pros" have, namely guaranteed byes at Grand Prix. Silver gets you 2 guaranteed byes and Gold/Platinum gets you 3. If we use the same 75% as above, but apply it to the 12 games they'd actually play and then add in the 3 guaranteed wins, that gives them a median of 12 wins, certainly enough for considerable consistency in succeeding. (if they have 2 byes, that gives a median of 11.75 wins)
I deliberately did not use GP as examples because it's hard to sort out the bye issue. My guess is that pros would do just fine without the byes but it's hard to prove. That's why I cited MTGO finishes. There are plenty of regulars whose consistent 5-0s can't be chalked up to luck, especially when it is with certain decks that benefit from luck much less than others.
For the best pro's having 3 byes is enormous. While the average competitor you'll face in the first 2-3 rounds is much lower than in the later rounds, modern is such a MU-dependent format and the range of decks being played on Day 1's Rounds 1-3 is so much wider than later in the tournament that the odds of you running up against a terrible MU increases significantly (unless of course you have terrible T1 MU's...). I can't stress enough how much of an advantage having 2 byes is, let alone 3! It's much much easier to go 10-2 dodging a wide field in the first 2-3 rounds than it is going 13-2 with no byes. It's also the main reason (imho) why you see the top players be more consistent on the GP circuit. I know very very good players who consistently put up 11-4 and 10-5 records with 0-1 byes, but they don't really grind GP's so don't get the necessary pro points or PWP for 2+ byes. I have a feeling WoTC does this because they know that the best players will hardly come T32 in 15+ round 2000+ player tournaments with no byes. GP's are all ready terribly poor EV, imagine having no byes as well. I doubt you'd see all the pros going to so many GP's without their 2 or 3 byes.
My point is, that a lot of people overplay the amount of luck in Modern, but underplay MU "luck" and the cumulative effect of luck for premier and GP level events that are 15+ rounds (especially with no byes). Or in other words, there is too much variance for even the best players in the world to consistently put up money finishes in GP level events without byes, which tells you a lot.
TL;DR The only way to know for sure whether "pros" are good enough to see consistently at the top tables at GP level events is to get rid of byes. They skew the data WAY too much.
Again (again), the Elo Project approach doesn't care about byes. You'd just see matches between players, the ratings of those players, and the outcome of their games. If you just check to see if those actual results are less/more/the same as the expected results based on rating, you can make a very good guess as to what kind of variance is at play. In chess, over a large number of games, a player who has +100 rating will generally beat the other player in about 64% of games. This actually plays out with real data too. According to the MTG Elo project, +200 points equates to a 60% increased win-rate. That scales with higher/lower ratings. Just check those numbers against the performance of difference players in Modern. If a 1900 player is only winning 50% of their matches over 1700 players, that means variance is probably accounting for about 10% of the difference (assuming N is big enough). I'm not doing this analysis, but I hope someone does.
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
Well, actually, 50% luck can still work with consistent player performance.
If 50% of the game is luck, that would mean that, logically, the other half is skill. Therefore, with sufficient skill, that remaining 50% would be completely filled, leaving you with 50% to fill that's luck. So you "start out" at 50%, and then would (randomly) get somewhere from 0-50% on the luck portion, averaging out to 25%. Put them together and you get 75%, meaning a median of 11.25 wins at a Grand Prix (15*0.75=11.25). A median of 11 out of 15 wins seems like enough to do well consistently.
And let's not forget certain advantages that the "pros" have, namely guaranteed byes at Grand Prix. Silver gets you 2 guaranteed byes and Gold/Platinum gets you 3. If we use the same 75% as above, but apply it to the 12 games they'd actually play and then add in the 3 guaranteed wins, that gives them a median of 12 wins, certainly enough for considerable consistency in succeeding. (if they have 2 byes, that gives a median of 11.75 wins)
I deliberately did not use GP as examples because it's hard to sort out the bye issue. My guess is that pros would do just fine without the byes but it's hard to prove. That's why I cited MTGO finishes. There are plenty of regulars whose consistent 5-0s can't be chalked up to luck, especially when it is with certain decks that benefit from luck much less than others.
For the best pro's having 3 byes is enormous. While the average competitor you'll face in the first 2-3 rounds is much lower than in the later rounds, modern is such a MU-dependent format and the range of decks being played on Day 1's Rounds 1-3 is so much wider than later in the tournament that the odds of you running up against a terrible MU increases significantly (unless of course you have terrible T1 MU's...). I can't stress enough how much of an advantage having 2 byes is, let alone 3! It's much much easier to go 10-2 dodging a wide field in the first 2-3 rounds than it is going 13-2 with no byes. It's also the main reason (imho) why you see the top players be more consistent on the GP circuit. I know very very good players who consistently put up 11-4 and 10-5 records with 0-1 byes, but they don't really grind GP's so don't get the necessary pro points or PWP for 2+ byes. I have a feeling WoTC does this because they know that the best players will hardly come T32 in 15+ round 2000+ player tournaments with no byes. GP's are all ready terribly poor EV, imagine having no byes as well. I doubt you'd see all the pros going to so many GP's without their 2 or 3 byes.
My point is, that a lot of people overplay the amount of luck in Modern, but underplay MU "luck" and the cumulative effect of luck for premier and GP level events that are 15+ rounds (especially with no byes). Or in other words, there is too much variance for even the best players in the world to consistently put up money finishes in GP level events without byes, which tells you a lot.
TL;DR The only way to know for sure whether "pros" are good enough to see consistently at the top tables at GP level events is to get rid of byes. They skew the data WAY too much.
Again (again), the Elo Project approach doesn't care about byes. You'd just see matches between players, the ratings of those players, and the outcome of their games. If you just check to see if those actual results are less/more/the same as the expected results based on rating, you can make a very good guess as to what kind of variance is at play. In chess, over a large number of games, a player who has +100 rating will generally beat the other player in about 64% of games. This actually plays out with real data too. According to the MTG Elo project, +200 points equates to a 60% increased win-rate. That scales with higher/lower ratings. Just check those numbers against the performance of difference players in Modern. If a 1900 player is only winning 50% of their matches over 1700 players, that means variance is probably accounting for about 10% of the difference (assuming N is big enough). I'm not doing this analysis, but I hope someone does.
ELO is a great tool when it comes to a game like Chess where variables are fixed and information is symmetrical so each game is a pure determinant of skill. In a game like Magic, and a format like Modern, ELO is a terrible metric to determine luck or variance because there are too many variables. Like I said before, people overplay luck in the abstract, but underplay MU luck in 15+ round tournaments where having 2 losses with one in the early rounds generally means you're out of contention, and 3 losses puts you in the bracket where you may not cash. Then you have to factor in how big an effect 2 or 3 byes is in the context of Day 1 fields and 15 rounds of play where early losses means your tie-breakers are absolutely atrocious. That's putting context to this conversation, where ELO has none. Is it "luck" in your definition when the 2200 player playing Jund loses to the Eld. Tron or Gx Tron player with 1500 rating? The "numbers" say that that should happen very rarely, but we know that the MU numbers say that it will play out in the 1500 rating players favor much more often than the 2200 Jund players favor. There's no way to parse this information because there is no deck data to go along with match results. You can't blindly factor luck due to the nature of the game. There are rough baseline guidelines when it comes to MU's, which need to be factored in first before luck is applied, which is then a factor of luck in itself (e.g. the MU lottery). It's much easier to build a deck to beat Day 2 expected decks and meta, than it is going in on Day 1 with no byes where the field is so much wider. That's the context that ELO cannot begin to approximate which you absolutely need to make a quantitative statement of how much luck is there in magic, especially in tournament contexts to gauge how successful players realistically can be/are in the parameters of the tournament.
ELO is a great tool when it comes to a game like Chess where variables are fixed and information is symmetrical so each game is a pure determinant of skill. In a game like Magic, and a format like Modern, ELO is a terrible metric to determine luck or variance because there are too many variables. Like I said before, people overplay luck in the abstract, but underplay MU luck in 15+ round tournaments where having 2 losses with one in the early rounds generally means you're out of contention, and 3 losses puts you in the bracket where you may not cash. Then you have to factor in how big an effect 2 or 3 byes is in the context of Day 1 fields and 15 rounds of play where early losses means your tie-breakers are absolutely atrocious. That's putting context to this conversation, where ELO has none. Is it "luck" in your definition when the 2200 player playing Jund loses to the Eld. Tron or Gx Tron player with 1500 rating? The "numbers" say that that should happen very rarely, but we know that the MU numbers say that it will play out in the 1500 rating players favor much more often than the 2200 Jund players favor. There's no way to parse this information because there is no deck data to go along with match results. You can't blindly factor luck due to the nature of the game. There are rough baseline guidelines when it comes to MU's, which need to be factored in first before luck is applied, which is then a factor of luck in itself (e.g. the MU lottery). It's much easier to build a deck to beat Day 2 expected decks and meta, than it is going in on Day 1 with no byes where the field is so much wider. That's the context that ELO cannot begin to approximate which you absolutely need to make a quantitative statement of how much luck is there in magic, especially in tournament contexts to gauge how successful players realistically can be/are in the parameters of the tournament.
Lots of misunderstandings in here. First of all, look at the ELO standings on the site. The leaderboard is full of top players, lines up very well with pro standings, and generally makes sense; whatever metric those guys used seems quite accurate, and they even explain their calculations on the site. They certainly have enough datapoints to calculate the rating, the math itself is solid, so I have little reason to doubt the accuracy of their system.
Second, this post again cites an absurdly specific example to illustrate a larger format principle. I don't care what factored into the 2200 player losing to the 1500 player in that single game. I just care how the 2200 players perform over a large dataset over the 1500 players; they should be favored to win 75% of those matches. That's because over a large dataset, we're going to encounter bad matchups, good matchups, and even matchups. Looking at enough players over enough tournaments, we're going to wash out the bad/good/even matchup split, unless large numbers of players are all playing the same bad decks with bad matchups. But we know from most Day 2 and Top 64 breakdowns that this isn't true and Modern is extremely diverse; if anything, more people are playing the decks with better matchups, so it will skew towards more wins than less.
Third, if you absolutely can't tolerate the site's ELO metric, just use the players' win percentages as rating indicators instead. Craig Wescoe's win rate is about 62% over all his tournaments captured on the site (870 games). Over all his Modern events since 2015, his win percentage is 61%, just below the average. This suggests to me that, in Wescoe's case alone, his overall skill is reflected in Modern and Modern variance/matchups aren't affecting it. But Josh McClain has a 62% overall win rate and a 71% Modern win rate, so he's probably doing something right in that format. If you repeated that breakdown for a large number of players, you would figure out to what extent overall win percentage was predictive of Modern win percentage.
People treat Magic statistics as an insanely complicated black box that no amount of analysis can crack. This viewpoint doesn't make sense to me. We have plenty of publicly available data to analyze and even the most basic metagame analyses tend to paint pictures that are mostly accurate. A much deeper dive into this ELO-based, player-level data would be even better. It would definitely help answer the question of this thread. I don't know if it would give a definitive answer. It probably wouldn't. But it would definitely point in the direction of an answer.
People treat Magic statistics as an insanely complicated black box that no amount of analysis can crack. This viewpoint doesn't make sense to me.
Oh it makes perfect sense. It's not rational, but it makes sense. It's just pride/ego. People like to think that they're much better at this game than they really are, which ends up perpetuating myths about luck and your black box of statistics.
People treat Magic statistics as an insanely complicated black box that no amount of analysis can crack. This viewpoint doesn't make sense to me.
Oh it makes perfect sense. It's not rational, but it makes sense. It's just pride/ego. People like to think that they're much better at this game than they really are, which ends up perpetuating myths about luck and your black box of statistics.
Totally agree. I think that accounts for a big chunk of the mythology.
[quote from="Aegraen »" url="http://www.mtgsalvation.com/forums/the-game/modern/779266-how-much-luck-is-involved-in-magic?comment=42"]I don't care what factored into the 2200 player losing to the 1500 player in that single game. I just care how the 2200 players perform over a large dataset over the 1500 players; they should be favored to win 75% of those matches.
Which means 25% they don't, but you want to say those 25% are weird corner cases and not indicative of the pervasive luck required throughout the game.
There are a lot of factors going into a game of Magic, and most of them are not probability based. Many of the factors that are probability based are mitigated by good deck building and the use of draw and filtering. It's pretty straightforward, really.
Which means 25% they don't, but you want to say those 25% are weird corner cases and not indicative of the pervasive luck required throughout the game.
If I could confirm that Magic games were 75% skill and 25% luck, that would make thia game only slightly less lucky than poker, which would disprove the hypotheses of many naysayers in this thread.
No one here believes MTG or Modern is purely skill. We all acknowledge luck as a factor. There's just a contingent of players who believe it decides most games, or at least enough games to account for why this contingent of players loses as many games as it does. This gets back to rc's point above, where a large subset of players think they are better than they actually are and can't imagine their losses are due to anything but variance. And yet, whenever we watch these players in actual games, we see far more games lost to bad decisions than variance. It's magnified in Modern because bad decisions are more costly than in a format like Standard; the power level is much higher here.
We see this same effect of variance overestimating in online ELO-based games. Players consistently believe they are hundreds of points better than they are and that their lower rating is due to bad allies and better opponents, i.e. matchmaking variance. But when you actually look at high level players and worse players, it's clear that skill carries far more games than matchmaking. This plays out in practice and in the stats, and Magic is the same.
Which means 25% they don't, but you want to say those 25% are weird corner cases and not indicative of the pervasive luck required throughout the game.
If I could confirm that Magic games were 75% skill and 25% luck, that would make thia game only slightly less lucky than poker, which would disprove the hypotheses of many naysayers in this thread.
No one here believes MTG or Modern is purely skill. We all acknowledge luck as a factor. There's just a contingent of players who believe it decides most games, or at least enough games to account for why this contingent of players loses as many games as it does. This gets back to rc's point above, where a large subset of players think they are better than they actually are and can't imagine their losses are due to anything but variance. And yet, whenever we watch these players in actual games, we see far more games lost to bad decisions than variance. It's magnified in Modern because bad decisions are more costly than in a format like Standard; the power level is much higher here.
Is 60/40 a good match up? Does this mean a good player should expect to lose 40% of the time, even when they have the better deck?
Which means 25% they don't, but you want to say those 25% are weird corner cases and not indicative of the pervasive luck required throughout the game.
If I could confirm that Magic games were 75% skill and 25% luck, that would make thia game only slightly less lucky than poker, which would disprove the hypotheses of many naysayers in this thread.
No one here believes MTG or Modern is purely skill. We all acknowledge luck as a factor. There's just a contingent of players who believe it decides most games, or at least enough games to account for why this contingent of players loses as many games as it does. This gets back to rc's point above, where a large subset of players think they are better than they actually are and can't imagine their losses are due to anything but variance. And yet, whenever we watch these players in actual games, we see far more games lost to bad decisions than variance. It's magnified in Modern because bad decisions are more costly than in a format like Standard; the power level is much higher here.
Is 60/40 a good match up? Does this mean a good player should expect to lose 40% of the time, even when they have the better deck?
One way to consider it: 60/40 means that a pilot with identical skill to his opponent will win about 60% of games in that matchup. But that isn't quite accurate. A more skilled opponent will probably bring that to 70%+. A less skilled opponent will bring it to 50% or worse. The issue here is that people are awful at estimating the true matchup %. They are either way too high (people think they are much better against Grixis DS than they are) or too low (people think some big mana matchups are more unwinnable than they are). They are also generally awful at estimating their own skill relative to the field. This means 60/40 as a win/loss % is super subjective and not very informative.
A better way to think of 60/40 is not winning 60% and losing 40% of games. It's that you are favored in 60% of games and unfavored in 40%. Skill will tip those margins.
Private Mod Note
():
Rollback Post to RevisionRollBack
Over-Extended/Modern Since 2010
To post a comment, please login or register a new account.
So your case and point is that you made an assumption about how a deck was built based on one card and because you happened to be right there must not be an element of luck involved? What if it wasn't using Gifts as the enabler, but some other card? I doubt we'd see you posting about the skill play of getting matched up against a rouge build?
He's going to be so mad when you're actually using Streetwraith to cycle through to a fast-mana combo Guild Feud or anything else. Lucky him, I guess?
That's cool and all, but the "luck" factor isn't your ability to guess correctly, it's whether or not your opponent actually has what you think he has. I have been Slaughter Games'd for some wildly irrelevant stuff (and total misses) when I was in my URx brewing phase. Based on the cards my opponent had seen, they usually made exactly the right call in what to name, but based on the luck of being matched up with me, instead of a more recognizable decklist, their SG result ranged from weak to whiff. Sometimes, even the "right" call is still wrong, and it's wrong because of luck.
UR ....... WUBR ........... WB ............. RGW ........ UBR ....... WUB .... BGU
Spells / Blink & Combo / Token Grind / Dino Tribal / Draw Cards / Zombies / Reanimate
I mean, I suppose that's going to be a luck-influenced miss in some very, very tiny % of games. In the vast majority of examples, however, when an opponent guesses Wraiths/Cliffs are most likely to be either Jund DS or Living End, they are mostly going to be correct. Or when they SG Snapcaster or Bolt, that's probably a hit in a huge majority of cases.
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
I mean, I suppose that's going to be a luck-influenced miss in some very, very tiny % of games. In the vast majority of examples, however, when an opponent guesses Wraiths/Cliffs are most likely to be either Jund DS or Living End, they are mostly going to be correct. Or when they SG Snapcaster or Bolt, that's probably a hit in a huge majority of cases.
This thread continues to illustrate one of the most annoying Magic, and Modern, behaviors. People take highly specific and absurd corner cases (Guild Feud combo??) and try to use them to support sweeping format generalizations. I read a thick layer of personal salt in some of these allegations, without any external support or even acknowledgement of argument weaknesses. Fot instance, the claim that Modern is 50%+ luck is absolutely ridiculous. There is no way we would see such consistent deck and player performance if that was true. There are also countless individual players, myself included, that can attest to win rates that are far more influenced by skill than luck. Just look at the consistent MTGO finishers. It's almost statistically impossible for some of those streaks to occur if winning in Modern was literally more luck-based than winning a coin toss.
Your example doesn't make a lot of sense. If your deck is way outside the norm, that isn't unlucky for your opponent, it's a good strategic play by you so long as your brew isn't total garbage that folds to commonly played decks.
All that said, I think your view on luck and Magic are ridiculously overblown. There is no way the data backs you up. The reason Owen Turtenwald won back to back GPs and made 7 top 8s in one year over 4 different formats was not because he was really lucky. It's because he was/is the best player in the world. If you or I played 10 matches against Owen where our decklist was favored 65/35, he would beat us 7 out of 10 times.
RBGLiving EndRBG
EDH
UFblthpU
BRXantchaRB
BGVarolzGB
URWZedruuWRU
If 50% of the game is luck, that would mean that, logically, the other half is skill. Therefore, with sufficient skill, that remaining 50% would be completely filled, leaving you with 50% to fill that's luck. So you "start out" at 50%, and then would (randomly) get somewhere from 0-50% on the luck portion, averaging out to 25%. Put them together and you get 75%, meaning a median of 11.25 wins at a Grand Prix (15*0.75=11.25). A median of 11 out of 15 wins seems like enough to do well consistently.
And let's not forget certain advantages that the "pros" have, namely guaranteed byes at Grand Prix. Silver gets you 2 guaranteed byes and Gold/Platinum gets you 3. If we use the same 75% as above, but apply it to the 12 games they'd actually play and then add in the 3 guaranteed wins, that gives them a median of 12 wins, certainly enough for considerable consistency in succeeding. (if they have 2 byes, that gives a median of 11.75 wins)
I deliberately did not use GP as examples because it's hard to sort out the bye issue. My guess is that pros would do just fine without the byes but it's hard to prove. That's why I cited MTGO finishes. There are plenty of regulars whose consistent 5-0s can't be chalked up to luck, especially when it is with certain decks that benefit from luck much less than others.
I think a lot of the skill of mtg is learning how to deal with the luck factor. If you have to mull to 6, you got unlucky, but knowing that you have to mull takes skill. You have plays where you should play around the last possible out for your opponent to turn a 95% win into a 100% win, but sometimes it is better to just pray they don't draw it and go in for the kill.
That would probably work but would also be unrealistically labor intensive. It would also fail to account for the complexities of real Modern or other format metagames.
All the data one needs to analyze this is on the MTG Elo Project website. Just see how often better players win/lose against worse players in Modern GP and see if that's higher, lower, or at the same rate as expected by their rating. Still labor intensive but it's all math and looking up players, not actually playing thousands of non-representative games.
Would you mind doing the legwork by any chance? It would be a fantastic article for this site and would help a lot of players better understand the nature of MTG.
I know I'm the OP and I should probably be the one to do it, but I honestly just don't have time with work and everything else.
Who will be the hero we need?
RWG Burn
GW Abzan Company
UR ....... WUBR ........... WB ............. RGW ........ UBR ....... WUB .... BGU
Spells / Blink & Combo / Token Grind / Dino Tribal / Draw Cards / Zombies / Reanimate
Like I said, the answer to this is on the MTG Elo project page. I even explained how to do it above. This will tell you based on real games whether or not players are winning in Modern as their rating predicts, or less/more than their rating predicts. If you have a bunch of players who are supposed to win 64% of their matches against the opponent based on rating, but then they are winning only 55% of matches, that's a pretty strong indicator that luck is at play in that other 9%. N is large enough, it's real world data, and it's literally all on the site. Someone just needs to crunch the numbers. I am not going to do it, but it's a great article/post/thread opportunity for anyone who wants to do the research and the math.
But again, it's not some unknowable element. It's very knowable and calculable. It will just be time consuming.
For the best pro's having 3 byes is enormous. While the average competitor you'll face in the first 2-3 rounds is much lower than in the later rounds, modern is such a MU-dependent format and the range of decks being played on Day 1's Rounds 1-3 is so much wider than later in the tournament that the odds of you running up against a terrible MU increases significantly (unless of course you have terrible T1 MU's...). I can't stress enough how much of an advantage having 2 byes is, let alone 3! It's much much easier to go 10-2 dodging a wide field in the first 2-3 rounds than it is going 13-2 with no byes. It's also the main reason (imho) why you see the top players be more consistent on the GP circuit. I know very very good players who consistently put up 11-4 and 10-5 records with 0-1 byes, but they don't really grind GP's so don't get the necessary pro points or PWP for 2+ byes. I have a feeling WoTC does this because they know that the best players will hardly come T32 in 15+ round 2000+ player tournaments with no byes. GP's are all ready terribly poor EV, imagine having no byes as well. I doubt you'd see all the pros going to so many GP's without their 2 or 3 byes.
My point is, that a lot of people overplay the amount of luck in Modern, but underplay MU "luck" and the cumulative effect of luck for premier and GP level events that are 15+ rounds (especially with no byes). Or in other words, there is too much variance for even the best players in the world to consistently put up money finishes in GP level events without byes, which tells you a lot.
TL;DR The only way to know for sure whether "pros" are good enough to see consistently at the top tables at GP level events is to get rid of byes. They skew the data WAY too much.
Again (again), the Elo Project approach doesn't care about byes. You'd just see matches between players, the ratings of those players, and the outcome of their games. If you just check to see if those actual results are less/more/the same as the expected results based on rating, you can make a very good guess as to what kind of variance is at play. In chess, over a large number of games, a player who has +100 rating will generally beat the other player in about 64% of games. This actually plays out with real data too. According to the MTG Elo project, +200 points equates to a 60% increased win-rate. That scales with higher/lower ratings. Just check those numbers against the performance of difference players in Modern. If a 1900 player is only winning 50% of their matches over 1700 players, that means variance is probably accounting for about 10% of the difference (assuming N is big enough). I'm not doing this analysis, but I hope someone does.
ELO is a great tool when it comes to a game like Chess where variables are fixed and information is symmetrical so each game is a pure determinant of skill. In a game like Magic, and a format like Modern, ELO is a terrible metric to determine luck or variance because there are too many variables. Like I said before, people overplay luck in the abstract, but underplay MU luck in 15+ round tournaments where having 2 losses with one in the early rounds generally means you're out of contention, and 3 losses puts you in the bracket where you may not cash. Then you have to factor in how big an effect 2 or 3 byes is in the context of Day 1 fields and 15 rounds of play where early losses means your tie-breakers are absolutely atrocious. That's putting context to this conversation, where ELO has none. Is it "luck" in your definition when the 2200 player playing Jund loses to the Eld. Tron or Gx Tron player with 1500 rating? The "numbers" say that that should happen very rarely, but we know that the MU numbers say that it will play out in the 1500 rating players favor much more often than the 2200 Jund players favor. There's no way to parse this information because there is no deck data to go along with match results. You can't blindly factor luck due to the nature of the game. There are rough baseline guidelines when it comes to MU's, which need to be factored in first before luck is applied, which is then a factor of luck in itself (e.g. the MU lottery). It's much easier to build a deck to beat Day 2 expected decks and meta, than it is going in on Day 1 with no byes where the field is so much wider. That's the context that ELO cannot begin to approximate which you absolutely need to make a quantitative statement of how much luck is there in magic, especially in tournament contexts to gauge how successful players realistically can be/are in the parameters of the tournament.
Lots of misunderstandings in here. First of all, look at the ELO standings on the site. The leaderboard is full of top players, lines up very well with pro standings, and generally makes sense; whatever metric those guys used seems quite accurate, and they even explain their calculations on the site. They certainly have enough datapoints to calculate the rating, the math itself is solid, so I have little reason to doubt the accuracy of their system.
Second, this post again cites an absurdly specific example to illustrate a larger format principle. I don't care what factored into the 2200 player losing to the 1500 player in that single game. I just care how the 2200 players perform over a large dataset over the 1500 players; they should be favored to win 75% of those matches. That's because over a large dataset, we're going to encounter bad matchups, good matchups, and even matchups. Looking at enough players over enough tournaments, we're going to wash out the bad/good/even matchup split, unless large numbers of players are all playing the same bad decks with bad matchups. But we know from most Day 2 and Top 64 breakdowns that this isn't true and Modern is extremely diverse; if anything, more people are playing the decks with better matchups, so it will skew towards more wins than less.
Third, if you absolutely can't tolerate the site's ELO metric, just use the players' win percentages as rating indicators instead. Craig Wescoe's win rate is about 62% over all his tournaments captured on the site (870 games). Over all his Modern events since 2015, his win percentage is 61%, just below the average. This suggests to me that, in Wescoe's case alone, his overall skill is reflected in Modern and Modern variance/matchups aren't affecting it. But Josh McClain has a 62% overall win rate and a 71% Modern win rate, so he's probably doing something right in that format. If you repeated that breakdown for a large number of players, you would figure out to what extent overall win percentage was predictive of Modern win percentage.
People treat Magic statistics as an insanely complicated black box that no amount of analysis can crack. This viewpoint doesn't make sense to me. We have plenty of publicly available data to analyze and even the most basic metagame analyses tend to paint pictures that are mostly accurate. A much deeper dive into this ELO-based, player-level data would be even better. It would definitely help answer the question of this thread. I don't know if it would give a definitive answer. It probably wouldn't. But it would definitely point in the direction of an answer.
Standard: lol no
Modern: BG/x, UR/x, Burn, Merfolk, Zoo, Storm
Legacy: Shardless BUG, Delver (BUG, RUG, Grixis), Landstill, Depths Combo, Merfolk
Vintage: Dark Times, BUG Fish, Merfolk
EDH: Teysa, Orzhov Scion / Krenko, Mob Boss / Stonebrow, Krosan Hero
Totally agree. I think that accounts for a big chunk of the mythology.
Which means 25% they don't, but you want to say those 25% are weird corner cases and not indicative of the pervasive luck required throughout the game.
CG
If I could confirm that Magic games were 75% skill and 25% luck, that would make thia game only slightly less lucky than poker, which would disprove the hypotheses of many naysayers in this thread.
No one here believes MTG or Modern is purely skill. We all acknowledge luck as a factor. There's just a contingent of players who believe it decides most games, or at least enough games to account for why this contingent of players loses as many games as it does. This gets back to rc's point above, where a large subset of players think they are better than they actually are and can't imagine their losses are due to anything but variance. And yet, whenever we watch these players in actual games, we see far more games lost to bad decisions than variance. It's magnified in Modern because bad decisions are more costly than in a format like Standard; the power level is much higher here.
We see this same effect of variance overestimating in online ELO-based games. Players consistently believe they are hundreds of points better than they are and that their lower rating is due to bad allies and better opponents, i.e. matchmaking variance. But when you actually look at high level players and worse players, it's clear that skill carries far more games than matchmaking. This plays out in practice and in the stats, and Magic is the same.
Is 60/40 a good match up? Does this mean a good player should expect to lose 40% of the time, even when they have the better deck?
CG
One way to consider it: 60/40 means that a pilot with identical skill to his opponent will win about 60% of games in that matchup. But that isn't quite accurate. A more skilled opponent will probably bring that to 70%+. A less skilled opponent will bring it to 50% or worse. The issue here is that people are awful at estimating the true matchup %. They are either way too high (people think they are much better against Grixis DS than they are) or too low (people think some big mana matchups are more unwinnable than they are). They are also generally awful at estimating their own skill relative to the field. This means 60/40 as a win/loss % is super subjective and not very informative.
A better way to think of 60/40 is not winning 60% and losing 40% of games. It's that you are favored in 60% of games and unfavored in 40%. Skill will tip those margins.