Brainstorm
Force of Will
Lion's Eye Diamond
Counterbalance
Sensei's Divining Top
Tarmogoyf
Phyrexian Dreadnaught
Goblin Lackey
Standstill
Natural Order
"For many readers, the biggest surprise to come out of the recent announcement was the decision to ban Sensei’s Divining Top in Extended. Making the decision to remove a card from an environment completely is never something taken lightly, and this time was no exception. Sensei's Divining Top caught the eye of Organized Play as being a potential problem during the Qualifier season for Pro Tour–Hollywood, but ultimately the decision was to monitor Top’s performance through the season and reconvene on the matter later in the year.
Ultimately Top 8s throughout the season were littered with the one-cost artifact either in conjunction with Counterbalance to lock opponents out of games, Trinket Mage to be found reliably, or (and usually in addition to) Onslaught’s sac-lands to allow players to shuffle away cards they didn’t wish to draw while peeking at a fresh set of three cards. Such a pervasive performance during a single season created a different problem as well: it made tournaments take too much time.
The constant activating of Divining Top bogs games down, which ultimately leads to an increase in the number of matches that go to time and beyond, which in turn leads to tournaments running much longer than they have historically. Furthermore, the Top encourages players to maximize the number of shuffle effects they play in a deck and the constant shuffling, cutting, presenting to an opponent to repeat the process, and then continuation of a turn exacerbated the situation. In the past the DCI has banned such cards on those grounds alone (Shahrazad is a good example of this, with Land Tax and Thawing Glaciers also having been banned for similar reasons) but in conjunction with the Top’s popularity during the last Extended PTQ season, the decision was to ban the card from the format it was harming."
Oh shit. Look what I did. Remember that now, and really hated that. It does bog a game down sometimes, but I don't think a card should be banned on that basis.
First, just for clarity, I never said anything about Ponder. I'm not nearly as concerned with the color makeup of the format as a lot of people in this thread are. I'm not - as I hope you know - in favor of banning Brainstorm yet. I do think that far fewer decks would be interested in playing Ponder + Preordain + Top (or Preordain + Top in the Ponder-banned case) in a Brainstorm and Dig-less world than you do though. Delver decks simply can't afford the tempo loss and Delver is a lot better with Dig banned. Shardless would rather use Sylvan Library because it's just a better card when you aren't trying to Miracle things and it's already tying up its mana almost every turn. Control decks (so in this discussion, UWx Blade) would certainly run it, but that's largely why I'd much rather see Dig gone than Top - I don't think that pure control can survive without CounterTop at this point. But I digress. Sean made a good point that right now, in the real world, if we're discussing banning Dig and Top to contain Omnitell and Miracles, we're making the same mistake with Brainstorm that was made with Necro Extended. I think the meta is pretty miserable to play in right now, but I'd rather see a cautious line from WotC of banning Dig and watching the meta closely for the next 3-6 months. If the meta adjusts and Miracles isn't dominant (and I think this will be the case), then all the reasonable people can go on with their lives and the diehards will still be calling for Brainstorm's head. If Miracles is the uncontested top deck in 3-6 months though, I think it's time to take a hard look at Brainstorm. If it comes to that, I'm honestly not sure what should be done anymore. Right now though, I think it's abundantly clear that Dig is broken and needs to go.
I got you the right way, don't worry. The point is that the classic Tempo shell is also close to unplayable if Brainstorm is banned because it suffers the same Problem as Miracles would: Running conditional cards (Daze, Stifle, Wasteland, Terminus, etc.) without a chance to switch them out. If you ban Brainstorm alone or Ponder alongside, doesn't matter that much as the Problem would most likely steer the Ux decks away from conditional stuff over time and narrowing playable choices in Ux further down the road.
The point I dislike about the Necropotence analogy is that Sean (and he hates Brainstorm as we all know, thus I'm not surprised the outcome was so unreflected) never asks if Brainstorm is the Necropotence or the Dark Ritual in this story. Isn't Brainstorm/Dark Ritual/Vengevine the card that fuels S&T/Necropotence/Survival to being such powerhouse-gameenders? The point Sean presents is too one-sided. How good is Necro w/o Ritual? How good is S&T/Delver/SFM w/o Braistorm? How good is Survival w/o Vengevine? And if we ban one of the two options, how would that affect the metagame? That are the questions to ask!
Edit: All you can do at this point is chopping the Hydras head and hope you don't start a downward spiral, like chopping Brainstorm to get hold of Miracles/S&T, then ban Ponder/Top/Preordain because they steered the format into a Blade mirror after the era of conditional cards ended with Brainstorm, then chop down combo, because running blue is so unattractive without cantrips and you can't keep combo in check w/o bannings, then you chop the cards in Aggro decks everyone runs as the format just narrowed down like Modern did and are back to just chopping the Hydras head once more
Edit 2: Need coffee. Still sleepy
www.theepicstorm.com - Your Source for The Epic Storm - Articles, Reports, Decktech and more!
Join us at Facebook!
Very much agree with this Lemn. I also think that the Necropotence analogy is flawed from mr. O'Brien. He seems on a crusade to get Brainstorm banned (I don't think he would disagree with that), and as I was listening to him on the Everyday Eternal Podcast he made the point that Necropotence would be played in a wide variety of archetypes (necro-control, necro-combo... necro-aggro(?)) The point was made to somehow invalidate the point that proponents of Brainstorm make, which is, that Brainstorm is not an all-dominating archtype but rather that Brainstorm fuels a wide variety of archtypes. I think necro, if unbanned, would at some point down the line establish what shell it would be most effective in - a combo variant I'm guessing - and there would be little point in playing necro-tempo or something else like that. I'm willing to concede that this is all just speculation of course
There was also the talk about "splashing" Necro into control decks which is total nonsense to do so, if Dark Ritual and Tendrils are legal in the format for a much more natural fit for Necro and there is exactly the point where the analogy is flawed: Necro would not power control or aggro these days. There is too much of an opportunity cost to run a BBB-costed card like Necro in regards to deckbuilding compared to Brainstorm for a single blue mana, which has also to be adresses in any discussion regarding the two cards, but that doesn't fit the bold black/white-picture Sean is painting.
www.theepicstorm.com - Your Source for The Epic Storm - Articles, Reports, Decktech and more!
Join us at Facebook!
Recording observations is not unscientific. What's unscientific is to assess a decks strength based entirely on it's success, while dismissing its failures at being not representative of the deck's position. That's exactly what you do if you look at top8s without consideration to representation. This is called cherry picking, and is the antithesis of science.
Except they make the same mistake you do! (BTW, Lands is 63%, MUD is 58%).
People look at top 8s and at win-rates and feel the numbers don't jive. How can MUD win 58% of its matches and Omnitell win only 47% when Omnitell gets so many more top showings? "Clearly" one of these measures is wrong. MUD may look good on paper, but Omni is actually winning events. Surely these spikes who are getting results know what they are doing, and besides, you cant argue with prize money! Right?
Wrong! The big mistake is thinking the win-rates and top8s are inconsistent, when in fact they are not. I'll take the example of Omnitell. This deck was roughly 20% of the field on day two of GP Kyoto (I don't have data for day one). The deck saw three spots in the top sixteen - which is also just under 20% of that bracket. This makes perfect sense with a 47% win-rate!
I didn't really need the stats for day two, though. If a break-even deck comprises ~20% of any given upper bracket, it's a good bet that deck represented ~20% of the field. Also vice-versa. This is the only way the numbers balance. Give me any two of a decks win-rate, it's percent of the field, or it's representation in a top bracket, and I can predict the other. There is variance of course - a 50/50 deck could in theory have all it's entrants winning exactly 50% of it's matches (no top eights), but as sample size increases, normal distribution prevails.
This isn't rocket science, but you'd swear it was. Some people really can't understand the maths - they cannot reconcile win-rates, so they toss out that data. But I think most people don't try, and stay willfully ignorant! Anyone clamoring for a DTT or BS ban (or SDT) will never admit that the decks they hate are not at all OP, but are propped up by sheer force of numbers. And people who love playing blue cantrips are invested in the belief that that they are playing a top deck. They would not want to have to switch decks or admit/accept that they are playing a deck because they like it, not because it is strong.
Whatever the reason, most people don't even want to look at win-rates! Just read these boards - the decks that are considered strong are those that win the most top8s. Period. So you can keep yourself ignorant by analyzing a fraction of the data, or you can look at all the data and get the full picture. It's all the same to me. I'd just like to make sure that new and impressionable players ca read an alternative analysis, and decide for themselves which pov is more reasonable.
This! Burn is the most consistent deck in Legacy, but it is not the best! Increased odds of drawing the card you want doesn't advance your record - winning matches does. Decks with higher win-rates consistently win matches more so than decks with lower win-rates.
Last edited by Crimhead; 06-05-2015 at 07:12 AM.
http://modernnexus.com/ (site is currently down) recently had an interesting article regarding the best deck of Modern. But instead of just giving out the win percentages, they also calculated the standard variation of said decks. Turns out that Bloom Titan has a win percentage of ~60% with an impressive standard derivation (0.006(%?) or something like that IIRC, I'll link the article later once the site is up again).
If win percentages stay high even after standard derivation, then yes, they're the best decks of the format. But before that, I'm not really impressed.
But a high standard deviation is what you want! Click the link I posted above on normal distributions. With a higher SD, you get more very good (aka, top 8) results and very bad results, while getting fewer average results.
SD doesn't alter an average (in this case, a win percentage) - the two are completely independent. SD measures the likelihood of any specific instance hitting the average, and the extent to which it deviates. In roulette, eg, betting on black or red has a lower SD than betting on a single inside number.
In an MTG tournament, if you deck has an averages win rate against the field of 50% or 60% (or whatever), you don't really want to win exactly that amount in any given tournament. You would rather have droughts and floods - a high SD! This matters less if you are choosing a feck for a smaller event, but still applies.
The problem is people don't understand math - even top players. That's why win-rates are not given the consideration they deserve, while top8 data is given blind faith despite being grossly stacked.
I admire your attempt to make a point with actual numbers in a comprehendable way, but you miss that you have no audience here, which really wants to understand what's behind those numbers. I once tried to link performance data and representation with actual numbers in this thread and set those into relation (basically what you tried to do with your last posts), which way called humbug because "only T8 results count". The next time you point at a diverse T8 with 3 non-blue deck, the same crowd yells "only T16 data count", and if you present a mixed T16, they look up the average percentage of Brainstorm for the month at mtgtop8.com and call Brainstorm overperforming (which I can proof wrong via the same set of data they picked) just to see the circle start all over. You can't ever win an argument here with math deeper than the front page numbers on mtgtop8.com and the like. You waste your time, therefore I have given up and moved straight to middlefinger waving in the face of ignorance for numbers
www.theepicstorm.com - Your Source for The Epic Storm - Articles, Reports, Decktech and more!
Join us at Facebook!
I'm kinda rusty in statistics after almost 10 years not having anything to do with it anymore.
Doesn't a high standard deviation only mean that the sample size was too small, thus resulting in a wider range of values where the true value lies within?
E.g. if Tom Ross has a win ratio 80% and Player B has a win ration of 40% with Infect, it averages out at 60%, but has a standard deviation of 60% (+/- 20%). Now if we had Player C with 55% and Player D had 65%, we would still have 60% win percentage on average, but standard deviation would be smaller - 60% (+/-14.6%). Tom Ross still took down the tournament and the win percentage stayed the same, but due to a bigger sample size, we're closer to the true value of the average win percentage.
The true average win percentage doesn't show us the individual win percentages of the players, though. A deck could very well be over- or underperforming based on sample size. Thing is, the higher the true average win percentage of a deck, the higher the chance of said deck placing (a deck with 50% wins average is less likely to place than a deck with 60% wins, for example).
While I agree that the 50% deck is going to place with more players by sheer numbers in the end due to normal distribution, I can't agree with just looking at win percentages and call it a day when we aren't sure what the true win percentage (and that is what we're looking for) actually is, especially if the sample size could be flawed. If the true average win percentage is higher, then it's the better performing deck. Maybe Lands does have a win percentage of 63% - maybe it has 58%, maybe 65%, but we can't say for sure until we know how accurate the value is.
That's why I'm calling for a deeper statistical analysis before making hasty conclusions.
As I did a PhD in econometrics I feel it's necessary to comment on your opinion here. In conventional probability theory we would model the relevant* win rate as a binary random variable X that can take the values of 0 (loss) and 1 (win). Let's neglect the fact that there are draws; their existence just complicates the math but doesn't change the result. Thus, the expected win rate is:
E(X) = P(X=1)
The expected variance of the win rate is:
Var(X) = E[X^2] - E(X)^2
= E(X) - E(X)^2 (--> this holds true because the win rate is a binary random variable)
= P(X=1)(1 - P(X=1))
Thus, the variance of the win rate is only dependent on its expected value. While this statement seems to support your argument at first it exposes it as rubbish at a closer look. Particularly, we can learn from the last equation above that the deck with the highest variance (0.25) will have a 50% win rate. The deck with a win rate of either 100% or 0% will have a variance of 0. The deck with a 75% win rate will have a variance of 0.21.
As a result, a "low variance win rate" is equal to either a high probability to win or a high probability to lose.
TL;DR: Only win rates count, variance or standard deviation are only dependent on win rates and can be completely disregarded if your approach to calculating your relevant win rate is correct.*
*All of the above holds true if you calculate the individual win rates against every possible deck in the field and weight these individual win rates by the expected probabilities of encountering each deck in the next tournament to arrive at the relevant win rate. This is the only correct approach to calculating the relevant win rate of an unknown match. For a "tournament win rate" you will have to start "Bernoulliing" using the calculated individual win rates and make assumptions about the decks you will face (can be extended to calculate a "top 8 rate").
Last edited by Ogh!; 06-05-2015 at 10:23 AM.
Except you use such stray results to justify the "health" of the format while completely ignoring regular results with 7-8/8 or 14-16/16 blue decks with Brainstorm placing, especially in big events like the GPs which have been a blue circlejerk in the past two years after Strasbourg (which had 4 non-Brainstorm decks placing).
When other people point out the numbers, you choose to ignore them while you call the format "diverse", how the high blue count is fine for Vintage (something we've already surpassed with Legacy by a good chunk by now - 67-71% vs 78-83%, depending on the sources you want to look at), how half of the format would die without Brainstorm and that everybody and their mother would quit without Brainstorm.
Don't try to argue with numbers when you're cherry-picking results to your liking and then try to call out people on their data.
You're right! I know this, just jad a brain fart this morning. It shows how easy it is to get this stuff wrong (and I'm someone with a background in math), and how counter-intuitive mathematical concepts can be, especially probability.
You do support my main argument; that a decks win-rate (if calculated properly) predicts it's chance of top 8ing.
I would never use such a vague term like "health" to describe a format. I soley talk about strategic diversity and leave the "colors matter" bullshit to others or the pointing at Brainstorm which is, for me, not different from running Fetchlands in every deck. I never said people would ragequit. I really hate if people put words in my mouth or if I have to point out obvious stuff like Miracles being unplayable as an archetype w/o Brainstorm, while certain users ignore the point I made about the conditional cards just for the sake of black/white picture painting
www.theepicstorm.com - Your Source for The Epic Storm - Articles, Reports, Decktech and more!
Join us at Facebook!
This thread is unhealthy at the best of times. (something I have been involved in causing, not of late I am thankful to say though.) But now we have people coming from offside to post exclusively in this one thread. I think it's not really helpful anymore, not that it really was in the first place.
And while I understand the argument that you leave the thread open to contain the bullshit, I feel like now the bullshit exists thanks to the thread. So this threads function has changed from containment to creation of shity conversation and manufactured issues. Putting a temp lock on this thread was the best thing that happened in recent history and I think it's time to put this shit to rest.
Lock this down, let's move on.
www.theepicstorm.com - Your Source for The Epic Storm - Articles, Reports, Decktech and more!
Join us at Facebook!
I agree that what you are explaining was the idea behind keeping this place open, what I disagree with is the idea that it is having a positive effect. We have people joining the site now just to post in this fucking place, the only reason they are here. we have the same old shit going round and round offering no positive effect (or any real effect) at all. We have the same people, making the same fucking arguments over and over again. This is not some toilet bowl collecting the shit, this is the bowels making it. People do not come here to the site to vent, if they did the thread I started and vented in in the off topic section would see more use.
This place, its just shity and getting worse. It is not offering a place to let out rage, it perpetuates it. It grants a place for people to feed not release. Also if you face a bad match, you talk about that in the thread in question. I am happy to vent in the Land thread about how I feel a match is bad, I am also happy to say I then get positive feedback on what I should do in the future. If I came here and tried that I would be attacked by something spouting shit about how I do not understand the "Health" of the format or some other bullshit response that means they do not have to actually think about what the hell they are saying.
I mean, lets examine the main shitty subject here, Brainstorm. Who here, among the name calling, shit slinging and otherwise bullshitiry has honestly had their views changed from what they have seen here? The point of a good debate is not to change the opinions of those you debate with, but of the audience whom is watching. The point in this thread is to stand on the mountain and piss down on everyone who does not agree with you, not to benefit anyone and change views. Its a circle jerk and cum is raining down all over the site and not for the better.
When I see comments made off site about The Source now, it is not about this is the place to come and talk about Legacy, it is about how there is a 500 page talk about Brainstorm. Not kidding. I have seen people state others should stay away because nothing productive happens here. This site is starting to be see as a home for this thread. This thing is gaining a life of its own and that is not a benefit for anyone. It needs to go, for the betterment of this place. Then if shit comes up later, you whack it on the head and whack the person who did it. I do not want to see the site become the brunt of jokes based only on the few morons in this thread that can not agree to disagree.
So you can say we have that audience, they have been watching, all they have seen is what the people who pop up now and again to say "How do I hide this thread" see and that ain't good. Time to get rid of this shit. If it aint going to be gotten rid of, it is time to at least make it viewable by Members only who have at least 10 posts or some shit and an opt out by the rest of us so we do not have to see it any more.
There are currently 3091 users browsing this thread. (0 members and 3091 guests)