The one thing that bothers me for quite a while now is magic skill difference, magic skill cap. I think the game doesn't give much space to excel. Recently I told my local magic community "any of us can win a world champion!". And this is so true and so wrong. Really, even my girldfriend can win Owen Turtenwald, LSV or enter_your_player_of_the_year. If she plays at GP she will play Burn. She doesn't play much because she doesn't care much and she is pretty bad (speaking of knowledge of modern decks). But even with month or two of training I think her chances are atleast 30-40% vs any great guy, vs world champion! Can you imagine this in other serious competition? That a new player with 6 month experience would have 30-40% vs world champion? In Starcraft, chess, World of Warcraft arena, kickboxing, figure skating? No way! You can argue that she used my skill for deck choice, but is there much skill needed to ask a friend or google for a best deck for beginner?
Other exmaple, I am above average player. When new set comes out, I can memorize LSV limited ratings, watch some draft videos, play some drafts and I'm set up. To have 25-40% at GP vs world champion!
There is one cool thing that measures player's performance: Elo Rating. 100 points difference means better player has 64% winrate vs lower player. 200 points - 76%, 300 - 85%, 400 - 91%. Player starts with 1600, it is zero-sum system, so average rating is 1600. But average tournament player is higher than that, according to my google research it is 1700. 1800-1900 is good and 1900-2000 is pro player, with high 19xx for best pros currently (and I seriously doubt 1950+ is actually achievable now as it requires fantastic winrate). So it is 2000-1600=400 points difference for average player vs best pro with best metagame call and luck upswing. That is skill cap measured in Elo points. Is this alot? No!
Let's take a look at other disciplines. Starcraft: there is no Elo ratings for casual players, pros start with rating of 2000, so average casual players' rating can be 1500-1600 (based on winrate vs pro). Best players have rating around 2450, thats 2450-1550=900 points difference, much bigger skill gap.
World of Warcraft: players start with rating of 1500, best players achive 3000 rating. But there is gear(swords, armor) factor: best players play only in best gear , low rating players can play in bad gear which hinders their performance. So assuming best gear gives around 200 points, difference is 3000-1700=1300.
Chess: average player is around 1300, top ten players have average of 2800. 1500 points difference! Do you remember what does it mean? 1800 guy basically destroys 1300 guy, 2300 guy destroys 1800 guy and 2800 guy destroys 2300 guy. Because skill matters.
Tl/dr: skill cap in magic is pretty low compared to other competitions and I want this to be changed.
There is one cool thing that measures player's performance: Elo Rating. 100 points difference means better player has 64% winrate vs lower player. 200 points - 76%, 300 - 85%, 400 - 91%. Player starts with 1600, it is zero-sum system, so average rating is 1600.
Hold it, buster. This is only true for ONE SPECIFIC SET OF K-VALUES. You are assuming too much, and probably ignoring that magic DID use Elo rating in the past. For over a decade, mind you. This system was abandoned in favor of the planeswalker points.
Anyway, to address your point, what you think is a bad thing is not necessarily so, and you are incorrect at estimating one's chances to win a MATCH, since you are focusing only on games. Your chance of winning a MATCH with a 25-40% win chance (let's say 33% for ease of calculations, shall we?) to win a game is only approximately 25.2% (Chance of winning game 1 and 2 = 11%, losing game one and winning 2 and 3 = 7.1%, winning game one and losing game two but winning game 3 = 7.1%). If you were looking at 25%, it would be 6.25% + 4.69% + 4.69% = 15.63% and if you were looking at 40%, you would end up with 16% + 9.6% + 9.6% = 35.2%.
Now, let's consider that the top players (LSV, PVDDR, Finkel) in terms of game win% have between 62% and 66% of game wins (official figures) against prepared players in the more demanding competitions (assuming that after a few rounds they are in the win brackets which is where the strongest players lie) and you see that the WELL PREPARED players have a 33% to 38% win chance. Realistically, someone who is not well prepared will have a much lower win chance.
Your post fails to take into consideration existing information, fails to have any actual meaningful calculations or data, and is nothing but a mindless rant. Please either do better, or keep these to a blog post or something.
Hagalaz
The thing you quoted has nothing to do with K-values, so I conclude you don't know what are you talking about. Read wiki and don't make you look like fool anymore please.
I was talking about match winrates, so your game winrate calculations don't apply. And does switch from 33% to 25% disproves my statement that my girlfriend can win world champion? No.
I mentioned skill difference for tournament players but seems like you didn't get it. Average tournament player - 1700 (above average player in general). Players in 'win brackers' - 18xx. LSV and PV having 64% winrate vs these players put them 100 points higher, to 19xx.
I provided arguments that Magic is much less skill based game than Stacraft, figure skating, wow or chess and so your point of my post being mindless rant disproved.
I believe it's not that the skill cap is low, what you're denouncing is the ratio between luck and skill in the game. Because the ratio is closer to a middle ground because of the huge amount of luck in Magic doesn't mean the skill cap isn't huge, in fact, probably much higher than the other games you mentioned. Skill just has less impact on the game. Now you can be for or against that, I have no opinion, just trying to clear out the facts.
It's also worth to mention that ELO rating was tried in MTG before the Planeswalker Points, and it stopped being used because it would encourage good ELO players to stay at home and avoid losing their precious points, even with decay, instead of having them come and play (read: buy more packs).
The thing you quoted has nothing to do with K-values, so I conclude you don't know what are you talking about. Read wiki and don't make you look like fool anymore please.
Really? So the expected win chance from a certain amount of rating difference is not related to how quickly rating can deviate? You ARE a fool if you don't understand this basic principle of bayesian rating systems in general and Elo in particular.
I was talking about match winrates
Then your numbers are completely unrealistic, and we need not go further.
I provided arguments that Magic is much less skill based game than Stacraft, figure skating, wow or chess and so your point of my post being mindless rant disproved.
Yes, and since I didn't address them, what can you conclude? You should conclude that this is correct. Starcraft and chess are much more skill based than magic. These are games with almost no variation. There is no need to discuss the obvious. However, this does not mean the rest of your post should be exempt from scrutiny.
Really. You said 'This is only true for ONE SPECIFIC SET OF K-VALUES'. Give me a real k-value counterexample at which statement you quoted becomes false. And which part of it do you question? There are a lot of sentences in quoted part
What numbers regarding match winrates are completely unrealistic? Prove me wrong.
What numbers regarding match winrates are completely unrealistic? Prove me wrong.
It's up to you to provide evidence for your own claims. "Prove me wrong" is not evidence. What statistics do you have to back the numbers that you've got here?
OP I think your viewpoint is common among young people who mostly play casually or in their local FNM/PPTQ circuit. I encourage you to go out and spike the nearest GP and make your way to Worlds from there, since it is so easy. With a 25-40% chance to win vs the best, it won't take you many tries at all to get it.
But seriously, its clear you haven't been playing too long if you didn't know MTG used to use elo; I encourage you to spend a bit more time playing before making judgments like his and claiming to be as good as a pro because you read articles and memorize what those pros tell you.
I believe it's not that the skill cap is low, what you're denouncing is the ratio between luck and skill in the game. Because the ratio is closer to a middle ground because of the huge amount of luck in Magic doesn't mean the skill cap isn't huge, in fact, probably much higher than the other games you mentioned. Skill just has less impact on the game. Now you can be for or against that, I have no opinion, just trying to clear out the facts.
It's also worth to mention that ELO rating was tried in MTG before the Planeswalker Points, and it stopped being used because it would encourage good ELO players to stay at home and avoid losing their precious points, even with decay, instead of having them come and play (read: buy more packs).
I agree I used term skillcap a bit wrong. I wanted to say best players' skill isn't enough to be close to unbeatable by average players. And there is no much room for pros to really improve their winrates.
What numbers regarding match winrates are completely unrealistic? Prove me wrong.
Where are you getting your "But even with month or two of training I think her chances are atleast 30-40% vs any great guy"?
And in a game with the variance of Magic, ELO ratings are bad. Because the game is not decided entirely by skill.
I had a friend that had a decent rating way back when. He lost one match due to a misplay, and then the next 2 to variance. His rating plummeted and it took him months to recover it.
Not because of skill - but because he just didn't draw lands. After that he swore off FNM and only played in larger events. Which is why ELO ratings for Magic just don't work - they get the people you want playing to stop playing.
I think your idea and the reality are not even close. I have been playing MTG for almost 23 years now. When I build a new deck I will goldfish it 200-300 times before I even play it against another player. I have been know to take it to bed and study every possible card combination in the deck etc. (pisses my wife off lol). Skill level between a kitchen table/LGS and a pro is a lot more than just a number. Red Deck Wins is probably the easiest deck to play properly. When you can get a 66% or higher win ratio with something like manaless dredge or no land vintage belcher, or successfully pilot legacy high tide combo then I will consider that a skilled player. I see this constantly at my store, newer players who get this idea they can compete at the pro level, they go spend all that money and scrub out in round 1 becasue they have no real idea of the skill involved at that level of competition. Yes Magic is a game of luck but proper deck construction and skill can suppress the luck factor by a lot.
In magic you are only as good as the best player you can beat constantly. If you have nothing but casuals and a few semi competitive players in your LGS then your skill level is going to be far too low to top 8 states or even most PTQ. Trust me I have gone undefeated in our local legacy scene for 35 games in a row, went to a Legacy tournament and went 2-2, and I play some of the best decks in legacy.
TL/DR
Your skill level is only as good as the best player you play.
There is one cool thing that measures player's performance: Elo Rating. 100 points difference means better player has 64% winrate vs lower player. 200 points - 76%, 300 - 85%, 400 - 91%. Player starts with 1600, it is zero-sum system, so average rating is 1600. But average tournament player is higher than that, according to my google research it is 1700. 1800-1900 is good and 1900-2000 is pro player, with high 19xx for best pros currently (and I seriously doubt 1950+ is actually achievable now as it requires fantastic winrate). So it is 2000-1600=400 points difference for average player vs best pro with best metagame call and luck upswing. That is skill cap measured in Elo points. Is this alot? No!
You realize that the DCI used to use an Elo rating system, right?
As for doubting 1950+ is actually achievable, think again.
Nicolai Herzog hit 2351 (Limited) in round 3 of the EU Fifth Dawn prerelease in Olso, although he dropped to 2341 by the end of the same event.
Herzog had 2350 (Limited) at the end of PT San Diego May 14, 2004. A month later, he completed a Limited event with a rating of 2334.
Jon Finkel holds fourth place for event-final Elo rating record (behind Herzog's 1st, 2nd, and 3rd place records) with a 2307 (Limited) from PT Kuala Lumpur.
No other player held an event-final rating above 2300, but Michael Turian had a 2325 (Limited) in round 16 of the same PT as Herzog's event-final 2350, and Gabriel Nassif had a 2318 (Constructed) in round 18 of PT Kobe the previous February.
The DCI discontinued the Elo system in favor of the current Planeswalker Points system because Elo discourages the best players from playing the game in order to protect their ratings, while the Planeswalker Points system practically forces them to play as much as possible.
There were two people at my local shop with limited ELO ratings above 1950. That level was not too difficult to achieve by most people at the pro level of play.
One thing that is really important is to note that major tournaments have many rounds for exactly this reason. While any individual match can go against a player just due to being unlucky, your odds of winning a 15 round GP with a 33% expected match win rate are approximately the following:
Probability of going 13-2: 0.00259%
Probability of going 14-1: 0.00018%
Probability of going 15-0: 0.0000059%
Probability of top 8: 0.0027759%
Then you need to go 3-0 in the top 8: 3.5937% chance
Now, in reality, your odds would be better than this because you don't have to play against nothing but pros for the entire time. You can play some of your rounds with a 50, 60, or 70% expected win rate because your opponents are even worse than you. But realistically, if you're at 33% vs the best, you are incredibly unlikely to take down a major tournament.
Well, yes, burn is an example of goldfishing with minimal skill required, but claiming all skill should be based on that is a bit absurd. Especially when getting into vintage you see that there is a lot of manipulation of the stack rules showing MTG mechanics skill, but often it will come down to whoever's goldfish activates first. Skillful play also isn't the be-all-end-all that you think it is, because in this format luck does indeed have a major role.
Remember the legendary [url=https://www.youtube.com/watch?v=ju_LZGBN5qU]Nassif Called Shot[/c], which was totally random chance. That's not skill, despite the fact that Nassif could probably tell you the chance of actually hitting his Ultimatium, it was still random chance. That kind of thing doesn't happen in world of warcraft or chess. Luck means a lot in MTG.
what OP seems to forget is that a Magic game starts WAY before you shuffle your deck and draw 7 cards.
your girlfriend (or any other first timer, for that matter) can learn to play any super aggro deck (burn included) that was previously assembled by a seasoned player and do well.
Modern is the worst format nowadays to prove anyone's skill at magic, as the metagame evolved to 2 players kinda playing solitaire magic trying to kill the other ASAP, it`s by far the least interactive of all the competitive formats.
So, yeah, a super aggro competitive modern deck could get some sweet results against the best players in the world even if the player behind it was the AI from Magic Duels in the "easy" settings.
that being said...
the game really starts way before that.
even if you don't brew your deck from scratch, choosing the right deck to play against a particular meta is one of the most challenging decisions you have to make in a game of magic.
you could say that Magic is 50% deckbuilding, 30% skill, 20% luck or something like that.
when you simply give a 100% ready and tested deck for someone without skill, you give her that 50% right away. Even if burn is not the best deck out there, it does quite a good job with just a small amount of luck.
OP I think your viewpoint is common among young people who mostly play casually or in their local FNM/PPTQ circuit. I encourage you to go out and spike the nearest GP and make your way to Worlds from there, since it is so easy. With a 25-40% chance to win vs the best, it won't take you many tries at all to get it.
But seriously, its clear you haven't been playing too long if you didn't know MTG used to use elo; I encourage you to spend a bit more time playing before making judgments like his and claiming to be as good as a pro because you read articles and memorize what those pros tell you.
Tough for him considering Top players have three byes, prestige and friends to get scooped into advanced rounds, etc.
Wow, seems like a whole thread against me)
I made a research on ratings prior to making this thread and I'm quite dissapointed I have to teach you the facts.
It's up to you to provide evidence for your own claims. "Prove me wrong" is not evidence. What statistics do you have to back the numbers that you've got here?
It's up to me to prove they are true
It's up to my opponent to prove they are wrong
Unless either is accomplished my statements are not true nor false, they are my unconfirmed opinion
Now I prove my statements:
Regarding my girlfriend having at least 30-40% with Burn - that was my estimate and community agrees with me http://www.mtgsalvation.com/forums/the-game/modern/704241-how-hard-it-is-to-play-burn
Regarding me having 25-40% in draft: with me having 1800 and my opponent having ~1950 Elo predicts I have 30% win chance.
The player mentioned in the first game will end up with a total of (starting rating + ( 10 * 10)) = 1600 + 100 = 1700 rating. The player mentioned in the second game will end up with a total rating of 1600 + (10 * 20) = 1800 rating. The first player has a 100 rating difference to the starting rating. The second player has 200 rating difference.
And yet, they have the exact same expected win% against players of the starting rating, simply because of the difference in K-value!
You use Elo algorithm wrong. After winning 10 points from opponent their rating difference becomes greater and subsequent wins will net less than 10 points.
By definition of Elo algorithm 1800 player has 76% expected winrate against 1600 player and 1700 player has 64% vs 1600, so your last statement is false and your whole argument is disproved.
Elo ratings were invented with variance in mind, so it handles luck well/perfectly and is good for magic (and players respect Elo ratings as quite accurate)
Wizards has already decided that it's not good for Magic. They want their pros playing more, not less. Elo encourages the pros to avoid playing in order to protect their score.
the problem with magic is that most decks play themselves,most is done while deckbuilding and not when piloting, there is always a best play /and there isnt much of a choice ,you play the value decks with more choices/branches can be made but i dont think they worth the hustle most of the time vs straightforward deck powerlevel (counter decks,pod-like,whateves)
this has turned me off magic ,while i still love lore/deckbuilding this is a huge issue.. that richard himself saw that's why he made another card game to "fix" this, netrunner (a medium deck with experienced pilot has great chances to win vs a poor pilot with a great deck)
if you have a great deck, player experience doesnt matter that much, it can make a difference, but in a few cases (stack and threat analysis,and be mindful of ca mostly)
rarely i have to think about the board state about what to do next.. and lose if i misjudge it.. maybe it's just me?
Elo ratings were invented with variance in mind, so it handles luck well/perfectly and is good for magic (and players respect Elo ratings as quite accurate)
Wizards has already decided that it's not good for Magic. They want their pros playing more, not less. Elo encourages the pros to avoid playing in order to protect their score.
They made a right thing making these ratings private. I was saying Elo ratings were good for measuring player skill.
Sure. There is a high variance/luck component in Magic, meaning a weak player has a non-negligible chance against the best....in a given game/match. But is that weak player going to make top 8 of a large tournament? The best comparison would be Poker. The best players there do consistently rise to the top...over time. But in a given hand/session the luck of the draw plays a large role in the outcome.
If you don't like, or just can't handle, the variance inherent to Magic, then I'd suggest picking up a different hobby to play competitively.
Sure. There is a high variance/luck component in Magic, meaning a weak player has a non-negligible chance against the best....in a given game/match. But is that weak player going to make top 8 of a large tournament? The best comparison would be Poker. The best players there do consistently rise to the top...over time. But in a given hand/session the luck of the draw plays a large role in the outcome.
If you don't like, or just can't handle, the variance inherent to Magic, then I'd suggest picking up a different hobby to play competitively.
Elo ratings were invented with variance in mind, so it handles luck well/perfectly and is good for magic (and players respect Elo ratings as quite accurate)
Wizards has already decided that it's not good for Magic. They want their pros playing more, not less. Elo encourages the pros to avoid playing in order to protect their score.
They made a right thing making these ratings private. I was saying Elo ratings were good for measuring player skill.
But when your rating is tied to byes in major tournaments, the system discourages competitive players from actually playing the game in ranked matches. That's exactly what Wizards doesn't want.
Variance and luck are inevitably tied to MTG. To change the variance would be to change the game, which i am sure none of us want. The only way to rectify your issue, reducing skill variance, would be to change the variance inherent in the game.
I want to clarify this 'wizards gave up Elo ratings because they were bad'.
Wizards used Elo wrong and this caused a lot of problems. Ratings were public so high-rated players screamed when lower rated player beat them because of luck, lower rated players were abused and humiliated. Wizards used huge k-values for no reason which made rating too shaky. So shaky that rating didn't reflect player's skill sometimes. They used ratings threshold to give players rewards (byes and qualifications), this made players to stop play and sit on their ratings.
Mainly ratings were abandoned because they incentivized people not to play and that was really stupid. Elo rating is great for tracking player's skill if reasonable k-value (like 20) is used. Elo ratings are treated as accurate (to some degree, minor fluctuations are natural) among pros and I provided links to prove this. Claims that Elo is bad for Magic because a lot of luck is involved are unfounded, Elo ratings were designed to take variance (luck) into account at its core. Elo and its time-decay successors are everywhere and universally acclaimed as best solutions to date.
To make ratings better Wizards need:
make k-value independent of event type. That use of k-values is stupid.
make k-value depend on number of total matches played by player. So new player can jump to its true rating faster.
make k-values reasonable: like 40 for new player (first 20 matches or so), 20 for others.
make top25 constructed and limited public so people know who are great
switch to time-decay system like Glicko so players were rewarded for playing frequently and sitting on rating was impossible (again, use reasonable parameters to not make system a failure)
You use Elo algorithm wrong. After winning 10 points from opponent their rating difference becomes greater and subsequent wins will net less than 10 points.
By definition of Elo algorithm 1800 player has 76% expected winrate against 1600 player and 1700 player has 64% vs 1600, so your last statement is false and your whole argument is disproved.
*FACEPALM*
Either you misunderstood my examples, or you do not understand how Elo works. I will attempt to diagnose which one it is with another question.
There are two matches in a tournament going on at the same time. The K-level for these matches is 32. On the first table, we have a 1800 rating player playing against a 1800 rating player. On the second table we have a 1500 rating player playing against a 1500 rating player. Player A wins the first table and player B wins the second table.
The question is: Which of the players gained more points from the match, A or B?
Please reply so I can know if you do understand Elo and there was a failure of communication in regards to the examples, or if you do not understand the system at all. Thanks.
Other exmaple, I am above average player. When new set comes out, I can memorize LSV limited ratings, watch some draft videos, play some drafts and I'm set up. To have 25-40% at GP vs world champion!
There is one cool thing that measures player's performance: Elo Rating. 100 points difference means better player has 64% winrate vs lower player. 200 points - 76%, 300 - 85%, 400 - 91%. Player starts with 1600, it is zero-sum system, so average rating is 1600. But average tournament player is higher than that, according to my google research it is 1700. 1800-1900 is good and 1900-2000 is pro player, with high 19xx for best pros currently (and I seriously doubt 1950+ is actually achievable now as it requires fantastic winrate). So it is 2000-1600=400 points difference for average player vs best pro with best metagame call and luck upswing. That is skill cap measured in Elo points. Is this alot? No!
Let's take a look at other disciplines. Starcraft: there is no Elo ratings for casual players, pros start with rating of 2000, so average casual players' rating can be 1500-1600 (based on winrate vs pro). Best players have rating around 2450, thats 2450-1550=900 points difference, much bigger skill gap.
World of Warcraft: players start with rating of 1500, best players achive 3000 rating. But there is gear(swords, armor) factor: best players play only in best gear , low rating players can play in bad gear which hinders their performance. So assuming best gear gives around 200 points, difference is 3000-1700=1300.
Chess: average player is around 1300, top ten players have average of 2800. 1500 points difference! Do you remember what does it mean? 1800 guy basically destroys 1300 guy, 2300 guy destroys 1800 guy and 2800 guy destroys 2300 guy. Because skill matters.
Tl/dr: skill cap in magic is pretty low compared to other competitions and I want this to be changed.
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!
I was talking about match winrates, so your game winrate calculations don't apply. And does switch from 33% to 25% disproves my statement that my girlfriend can win world champion? No.
I mentioned skill difference for tournament players but seems like you didn't get it. Average tournament player - 1700 (above average player in general). Players in 'win brackers' - 18xx. LSV and PV having 64% winrate vs these players put them 100 points higher, to 19xx.
I provided arguments that Magic is much less skill based game than Stacraft, figure skating, wow or chess and so your point of my post being mindless rant disproved.
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!
It's also worth to mention that ELO rating was tried in MTG before the Planeswalker Points, and it stopped being used because it would encourage good ELO players to stay at home and avoid losing their precious points, even with decay, instead of having them come and play (read: buy more packs).
What numbers regarding match winrates are completely unrealistic? Prove me wrong.
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!
It's up to you to provide evidence for your own claims. "Prove me wrong" is not evidence. What statistics do you have to back the numbers that you've got here?
But seriously, its clear you haven't been playing too long if you didn't know MTG used to use elo; I encourage you to spend a bit more time playing before making judgments like his and claiming to be as good as a pro because you read articles and memorize what those pros tell you.
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!
Where are you getting your "But even with month or two of training I think her chances are atleast 30-40% vs any great guy"?
And in a game with the variance of Magic, ELO ratings are bad. Because the game is not decided entirely by skill.
I had a friend that had a decent rating way back when. He lost one match due to a misplay, and then the next 2 to variance. His rating plummeted and it took him months to recover it.
Not because of skill - but because he just didn't draw lands. After that he swore off FNM and only played in larger events. Which is why ELO ratings for Magic just don't work - they get the people you want playing to stop playing.
In magic you are only as good as the best player you can beat constantly. If you have nothing but casuals and a few semi competitive players in your LGS then your skill level is going to be far too low to top 8 states or even most PTQ. Trust me I have gone undefeated in our local legacy scene for 35 games in a row, went to a Legacy tournament and went 2-2, and I play some of the best decks in legacy.
TL/DR
Your skill level is only as good as the best player you play.
As for doubting 1950+ is actually achievable, think again.
Nicolai Herzog hit 2351 (Limited) in round 3 of the EU Fifth Dawn prerelease in Olso, although he dropped to 2341 by the end of the same event.
Herzog had 2350 (Limited) at the end of PT San Diego May 14, 2004. A month later, he completed a Limited event with a rating of 2334.
Jon Finkel holds fourth place for event-final Elo rating record (behind Herzog's 1st, 2nd, and 3rd place records) with a 2307 (Limited) from PT Kuala Lumpur.
No other player held an event-final rating above 2300, but Michael Turian had a 2325 (Limited) in round 16 of the same PT as Herzog's event-final 2350, and Gabriel Nassif had a 2318 (Constructed) in round 18 of PT Kobe the previous February.
The DCI discontinued the Elo system in favor of the current Planeswalker Points system because Elo discourages the best players from playing the game in order to protect their ratings, while the Planeswalker Points system practically forces them to play as much as possible.
Two Score, Minus Two or: A Stargate Tail
(Image by totallynotabrony)
One thing that is really important is to note that major tournaments have many rounds for exactly this reason. While any individual match can go against a player just due to being unlucky, your odds of winning a 15 round GP with a 33% expected match win rate are approximately the following:
Probability of going 13-2: 0.00259%
Probability of going 14-1: 0.00018%
Probability of going 15-0: 0.0000059%
Probability of top 8: 0.0027759%
Then you need to go 3-0 in the top 8: 3.5937% chance
Now, in reality, your odds would be better than this because you don't have to play against nothing but pros for the entire time. You can play some of your rounds with a 50, 60, or 70% expected win rate because your opponents are even worse than you. But realistically, if you're at 33% vs the best, you are incredibly unlikely to take down a major tournament.
Remember the legendary [url=https://www.youtube.com/watch?v=ju_LZGBN5qU]Nassif Called Shot[/c], which was totally random chance. That's not skill, despite the fact that Nassif could probably tell you the chance of actually hitting his Ultimatium, it was still random chance. That kind of thing doesn't happen in world of warcraft or chess. Luck means a lot in MTG.
your girlfriend (or any other first timer, for that matter) can learn to play any super aggro deck (burn included) that was previously assembled by a seasoned player and do well.
Modern is the worst format nowadays to prove anyone's skill at magic, as the metagame evolved to 2 players kinda playing solitaire magic trying to kill the other ASAP, it`s by far the least interactive of all the competitive formats.
So, yeah, a super aggro competitive modern deck could get some sweet results against the best players in the world even if the player behind it was the AI from Magic Duels in the "easy" settings.
that being said...
the game really starts way before that.
even if you don't brew your deck from scratch, choosing the right deck to play against a particular meta is one of the most challenging decisions you have to make in a game of magic.
you could say that Magic is 50% deckbuilding, 30% skill, 20% luck or something like that.
when you simply give a 100% ready and tested deck for someone without skill, you give her that 50% right away. Even if burn is not the best deck out there, it does quite a good job with just a small amount of luck.
Tough for him considering Top players have three byes, prestige and friends to get scooped into advanced rounds, etc.
I made a research on ratings prior to making this thread and I'm quite dissapointed I have to teach you the facts.
I know that mtgo has Elo ratings, I even told you the distribution of average/good/best players' ratings.
Elo ratings were invented with variance in mind, so it handles luck well/perfectly and is good for magic (and players respect Elo ratings as quite accurate)
Elo ratings of 2000++ were in era of huge k-values. K-value determines the speed of rating change after a match. These k-values were used improperly by wizards and thus players' ratings did fluctuate a lot, big swings after a win or loss streak.
On a realistic scale a player with a 2000 rating can easily be a player of the year.
Pros to back me up:
https://twitter.com/Juzam_/status/396305895916728321
https://www.reddit.com/r/lrcast/comments/2jhahs/elo_rating_coming_back_to_mtgo_yay_or_nay/clbxpwu
http://www.starcitygames.com/php/news/print.php?Article=21850
It's up to me to prove they are true
It's up to my opponent to prove they are wrong
Unless either is accomplished my statements are not true nor false, they are my unconfirmed opinion
Now I prove my statements:
Regarding my girlfriend having at least 30-40% with Burn - that was my estimate and community agrees with me http://www.mtgsalvation.com/forums/the-game/modern/704241-how-hard-it-is-to-play-burn
Regarding me having 25-40% in draft: with me having 1800 and my opponent having ~1950 Elo predicts I have 30% win chance.
You use Elo algorithm wrong. After winning 10 points from opponent their rating difference becomes greater and subsequent wins will net less than 10 points.
By definition of Elo algorithm 1800 player has 76% expected winrate against 1600 player and 1700 player has 64% vs 1600, so your last statement is false and your whole argument is disproved.
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!
That's not how the burden of proof works.
Two Score, Minus Two or: A Stargate Tail
(Image by totallynotabrony)
this has turned me off magic ,while i still love lore/deckbuilding this is a huge issue.. that richard himself saw that's why he made another card game to "fix" this, netrunner (a medium deck with experienced pilot has great chances to win vs a poor pilot with a great deck)
if you have a great deck, player experience doesnt matter that much, it can make a difference, but in a few cases (stack and threat analysis,and be mindful of ca mostly)
rarely i have to think about the board state about what to do next.. and lose if i misjudge it.. maybe it's just me?
Δε φοβάμαι τίποτα...
Είμαι Άνεργος.
Grimstringer on Cockatrice, add me if you wanna
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!
If you don't like, or just can't handle, the variance inherent to Magic, then I'd suggest picking up a different hobby to play competitively.
If you don't like, or just can't handle, the variance inherent to Magic, then I'd suggest picking up a different hobby to play competitively.
Two Score, Minus Two or: A Stargate Tail
(Image by totallynotabrony)
Wizards used Elo wrong and this caused a lot of problems. Ratings were public so high-rated players screamed when lower rated player beat them because of luck, lower rated players were abused and humiliated. Wizards used huge k-values for no reason which made rating too shaky. So shaky that rating didn't reflect player's skill sometimes. They used ratings threshold to give players rewards (byes and qualifications), this made players to stop play and sit on their ratings.
Mainly ratings were abandoned because they incentivized people not to play and that was really stupid. Elo rating is great for tracking player's skill if reasonable k-value (like 20) is used. Elo ratings are treated as accurate (to some degree, minor fluctuations are natural) among pros and I provided links to prove this. Claims that Elo is bad for Magic because a lot of luck is involved are unfounded, Elo ratings were designed to take variance (luck) into account at its core. Elo and its time-decay successors are everywhere and universally acclaimed as best solutions to date.
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!
make k-value independent of event type. That use of k-values is stupid.
make k-value depend on number of total matches played by player. So new player can jump to its true rating faster.
make k-values reasonable: like 40 for new player (first 20 matches or so), 20 for others.
make top25 constructed and limited public so people know who are great
switch to time-decay system like Glicko so players were rewarded for playing frequently and sitting on rating was impossible (again, use reasonable parameters to not make system a failure)
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!
G Green Stompy
RG Shamans
UB Mill
UG Infect
WUBRG Slivers!