[Deck] UW Tempo

[Forbiddian]

The main thing here is that you're pairing player that are really good with UWT (the creators of the deck and people that play it non stop). Merfolk, on the other hand, is a deck that is played by a lot more people. Especially when looking at your videos it doesn't exactly seem like you're playing against the most competent players out there. Doesn't mean that this will make a difference in the actual statistics, does mean that if you actually want to prove something you need to play against people that have a clue about what they're doing.

What's funny is I'm actually playing against one of the top UWT players who is definitely counting toward the data and I think three people commented about how bad he was.

There's no such thing as perfect data, but my analyses have all been spot on and made UWT a profitable deck in the SCGs 5ks and a constant presence in the top 8/top 16 at every SCGs despite few people playing it. The results that we've seen in SCGs, in tournaments around the world reflect exactly the predictions that I've made.

If you still don't believe me, that's your problem. If you still want to nitpick my imperfect data set because I'm not Yahweh Fucking Elohim, be my guest, but I've already done enough to use the absolute worst possible data that I could and guess what: UW Tempo is still better than Zur.

[Nessaja]

What's funny is I'm actually playing against one of the top UWT players who is definitely counting toward the data and I think three people commented about how bad he was.

But a top UWT player doesn't equal a top Merfolk player. When you look at your video you can easily spot a lot of mistakes. Getting good data is about taking variables out, what you have currently proven is that you consistently win against players that (on video) don't seem to be any good at playing Merfolk. Why don't you just play against someone from the merfolk topic (there have been plenty of suggestions for that), video the set of games, and get the thing over with? Put good players against good players.

[Forbiddian]

But a top UWT player doesn't equal a top Merfolk player. When you look at your video you can easily spot a lot of mistakes. Getting good data is about taking variables out, what you have currently proven is that you consistently win against players that (on video) don't seem to be any good at playing Merfolk. Why don't you just play against someone from the merfolk topic (there have been plenty of suggestions for that), video the set of games, and get the thing over with? Put good players against good players.

I mean, this is where the arguments go: It's down to some pathetic red herring. "There was this one time where you played this one guy in some exhibition games months ago. I know you don't even count those games as testing results, but still it magically invalidates your testing results." Listen to yourself. It's not even an argument anymore, it's just a misstatement of fact.

The games were just six games that show how to play against Merfolk, which strategies to use and how to take advantage of different opportunities.

I mean, you're smarter than that, right? God I hope so. For the videos, I usually only pick games that are close or where something really interesting happens (already obviously not normal). I usually don't cast games where I or my opponent mulligans to five or games where my opponent resigns randomly, although I tried to have some kind of video up for each archetype. The videos are partly for entertainment and partly for education in terms of how to play UWT, but I really hope people don't go through the videos and like count up my wins/losses and then think that somehow reflects UWT's actual win/loss ratio since the videos are definitely edited/selected for interesting value. And even of the interesting games, it's better to pick a win than a loss.

I mean, they're highlights. It'd be like watching the Top 10 Plays of the Week and then thinking that offenses score on every possession. I really didn't think I needed to point this out, but I guess some people are as credulous as christians in an easter egg hunt.


And @ your suggestion: I'm not the person dodging. I sent out a bunch of PMs a while ago to play matches, and nobody was interested.

[Nessaja]

I mean, you're smarter than that, right? God I hope so. For the videos, I usually only pick games that are close or where something really interesting happens (already obviously not normal). I usually don't cast games where I or my opponent mulligans to five or games where my opponent resigns randomly, although I tried to have some kind of video up for each archetype. The videos are partly for entertainment and partly for education in terms of how to play UWT, but I really hope people don't go through the videos and like count up my wins/losses and then think that somehow reflects UWT's actual win/loss ratio since the videos are definitely edited/selected for interesting value. And even of the interesting games, it's better to pick a win than a loss.

Point is that your opponents were really terrible in those videos, I don't see how that is either interesting or a "highlight" but you don't learn a lot from bad opponents (IMO). If you play games where your opponents aren't as bad you should just upload those. I mean; you already got the tools, you already make videos why don't you upload videos where you play against better opponents?

Naturally I didn't count your wins and losses from the videos, surely you are smarter then thinking that.

It's down to some pathetic red herring. "There was this one time where you played this one guy in some exhibition games months ago. I know you don't even count those games as testing results, but still it magically invalidates your testing results." Listen to yourself. It's not even an argument anymore, it's just a misstatement of fact.

False.
And for future reference; before you call someone out on "misstatement of fact" you should at the very least be sure what he or she is talking about. You can't put someone words in the mouth and then proceed to call them out on false logic when it's your internal dialogue.

What I say is that if others cannot replicate your results you can proof your point with video material and the current video material does not suffice.

With the current knowledge you have spread in the world there are several people who cannot replicate your results, seeing is believing, as such my suggestion.

[(nameless one)]

to Forbiddian and pi4meterftw:

I currently do not have any Tundras. Would your current list work without any Tundras? Possibly replace them with Hallowed Fountains or even with just Basic lands and rely on fetchlands for a balanced manabase.

Any additional advices for that?

Thanks

[Tinefol]

Went 2-2-1 today,
R1: BGW HelmLine (0-2), he quickly comboed out me game1, and in game2 I incorrectly sideboarded and had fallen as a victim of his aggro plan. Got flooded too, but that doesn't excuse sloppiness on my part.
R2: Team America (1-2). A horrible mistake game1 prevented me from taking control f the board, game2 went rather smoothly, and in game3 I got really flooded - 15/17 lands. I didn't draw into Wayfarer in any of the games.
R3: Draw 1-1-1 with Food Chain combo, where I should've lost, if not for the turns. Emrakul is the bitch.
R4: Won against White Weenie 2-0
R5: Won against mono black aggro 2-0.

Finished 10th and I have only myself to blame, I still blunder a lot.

[steveo356]

This list will not work without tundras. you have to play and replay your lands several times which makes shock duals absolute trash. I mean it is not uncommon at all between daze and fathom seer to life up the same 2 tundras for then entirety of the game.
if you could get your mits on even 2 tundras you may be able to make the deck work

[Grollub]

Also, 9-0-2 (if ~9-1) is 1% likely at p=0.5. It's small, but still big enough to prove at the 99% confidence interval that UWT beats Merfolk.

I'm neutral in the feud of UWT vs. Merfolk, but please don't misuse statistics...

[Forbiddian]

I'm neutral in the feud of UWT vs. Merfolk, but please don't misuse statistics...

What's wrong with that?

The alternative, I guess, would be an Elo analysis (which can use draws), but there aren't good probability models for random drift on Elo.

But the assumption that 2 draws is the same as 0-1-0 is if anything a conservative estimate of performance. I mean, for example if you're 1-1-8, it's not the same as being 1-5. But anyway, for the statistics, 9 Wins in 10 Matches if the matchup is really even is only 1% probable. That means we're 99% confident that UW Tempo beats Merfolk. That's what a confidence interval means, what a null hypothesis is....

Anyway, if I did something wrong, let me know, but it seems like maybe you need to take a statistics class.

[Grollub]

What's wrong with that?

Because correlation does not imply causation. ;-)

The model you're working on is way to complex to use such a small sample, and there's a severe lack of testing correalation between various variables (player and deck version most noticable). Without these tests, the interoperation of your data could just as easily be twisted into, "UWT beat Merfolk during winter (or whatever few couple of months the data cover)" - which obviously is absurd, as Magic as far as I'm concerned doesn't care if it's winter, spring, summer or fall.
EDIT: Ah, bad example as Magic does care - new sets. Assume it had rained in all the tournaments, and change "during winter" to "during rainy days".

How'd you apply an Elo system when there's only two contestants in the analysis? Or is there's another Elo analysis, than the ranking system, I'm unaware of?

That aside!
After considering it (this is purely in theory, as I haven't tried the match up first hand), I'd say UWT should have a considerable advantage in the Merfolk match-up, there's just so many tools to foil Merfolks particular gameplan it's silly.

@Seasoned UWT players: I know this is probably not a very relevant match-up, but how's does it fare versus mono-red Sligh (it's beginning to pop up quite regular at my FNM...)? SB solution(s) - Chill and Absolute Law to go with the Tutor plan?

[Forbiddian]

Because correlation does not imply causation. ;-)

Not sure if you're kidding because of the wink, but your entire post was an incredible stretch.

But a binomial distribution doesn't care about the complexity of the game, as long as you end up with a binary result. My group had two draws, so I assumed that's where you're correcting me (though I did treat them conservatively). Binary distributions are routinely used to solve problems as complicated as coin flips (honestly: think of the physics of a coin flip), heads up poker matches, chess games, I don't really see why you'd consider chess to be different.

But in general, I don't know how you got this idea that Magic is "too complicated" to use statistics on. It's reminiscent of people saying that life is too complex to have evolved. And incidentally, you can still get an Elo rating with only two agents, though it's pretty simple it doesn't increase without bound. The problem is I don't know of a way to account for random drift on Elo.

There's another rating system I think called Gechko or something that tries, but it's pretty complicated and I didn't take enough statistics classes.



@ Other point: Vs. Mono Red Sligh, keep Burrenton Forge-Tender in the sideboard. Hoses the hell out of anything Red.

I wouldn't recommend Chill, the standard gameplan is to out card-advantage them. They can't kill you with their opening hand, so the goal is to just trade cards away. Each time you stop one of their cards, you more than timewalk them, since they need to draw into another damage spell to kill you off. Hitting them in the card count is by far the most effective way to safely slow them down. Chill slows them down, but they're in the end still able to dump their hand and Chill isn't that useful.

Although it does have some cool side applications if Red Storm is very popular in your metagame, it's not a good hate spell for Sligh or Burn.

I'd recommend stuff that trades with their cards, like I dunno... more spell pierces, Renewed Faith, I think there was some life gain card printed in the new set that's pretty good.

[steveo356]

forbiddian pretty muh hit it on the head and i havent had any trouble playing against it and i beat it twice at weekly legay this week. MoM>any creature they an play. I dont have forgetender sided any more i just bring in propaganda and a pair of ET because jitte is most often the card that will win this match for you 9/10 games

[Oiolosse]

yeah, mtg is not too complicated to do statistics on. Compare with quantum mechanics. It would make for a fun project though, applying what we learn in stat to mtg. I mean really applying it.

[Octopusman]

Ordered quantity 4 mirage Enlightened Tutors like I said (auctions were listed as x1, but you can enter 4 and 4 were available).
Only 3 came. It was from a pretty reputable seller so I'm not that worried even though I still have gotten no response from the emails I sent.
Sucks balls. And I have never had an issue getting cards from reputable ebay sellers until now.


Anyway, I should have some opportunities to test this out West in the relatively small San Francisco Bay Area meta.
The semester is ending so I'll have an opportunity to play in San Jose and they are usually small 14+ people weekly tourneys. There may be tourney in Berkeley as well.

There have been some really nice write ups recently but I think for the sake of testing variations of the deck (perhaps ill-conceived) I'll try to run something slightly different.
With that said, there's no sense in trying anything mentioned previously. If I can't come up with anything I will just test the proven list.

I also mentioned that I'm a longtime (non-merfolk 'god can't believe you have to caveat it this way these days') Fish player. I think there may be some cards that have become "outclassed" that could work well in UW tempo. I say "outclassed" because not only because of power creep, but because overtime things fall out of favor/fashion. Sure, most of the time they are replaced because through testing people find out that other cards are better. Othertimes the meta evolves, or a deck is created that, because of its design, can "turn on" cards that may need to be rethought.
Yes, I think curiosity is not good for legacy.

Here are some Fishish lists. Some old, some new (some T1):
http://www.starcitygames.com/php/new...cle/10025.html
http://www.magic-league.com/deck/27299/legacy_t15.html
http://forums.mtgsalvation.com/showthread.php?p=5296466
http://www.magic-league.com/deck/59208/ant.html
http://magicleague.startlogic.com/de..._affinity.html
and the classic
http://www.mtgthesource.com/forums/s...ck%5D-UWb-Fish

Just to get some ideas! I really like seeing this thread updated. Like others have mentioned, if the deck is going to stay the same or go anywhere it all boils down to testing testing testing and sharing results/MU. I'm stoked to see how others do.

I think the reports are awesome. Keep it up.

[jazzykat]

I don't know if anyone has shared my sentiments but I think it is very cool that a control-aggro deck w/o tarmogoyf that doesn't lose to itself was developed, shared, and taught to tne community.

Depending on how the videos were done you can still learn the desired lesson from a bad game.

[pi4meterftw]

Are you seriously still doing this? Stop pretending any self reported data is non-biased and I'll stop trolling your thread. Not to mention you have failed to include at least 3 less than 50:50 performances (from "credible" sources) that I've seen reported here, meaning that your biased reporting results are biased by your selection.

If we included reports corresponding to all of your whining, then that would definitely be biased data since you only whine when we overlook a bad report. Your contributions as of late to this thread seem to be incessantly begging us to have perfect data or whatever, even though we don't even claim to have perfect data.

Didn't Matt reveal tournament results at the SCG 5Ks for all UWT participants? I can't even imagine a plausible way to make that sound like it's biased. Why don't you just go off that if you're so particular?

Or do you think like: SCG 5Ks favor UWT players because that's an explanation that's consistent with the data and serves the platform I'm advancing! :):):):):):):)

[Grollub]

@Regarding red: Sounds good! I'm looking forward to try this badboy out in the soon future, looks like a blast to play.

Is the decklist on first post up to date with the latest tweaks?

But in general, I don't know how you got this idea that Magic is "too complicated" to use statistics on. It's reminiscent of people saying that life is too complex to have evolved.

The issue isn't that you cannot apply a binomial distribution or statistics in general, the issue is the variables you're discarding by only looking at such a simple model. Would you for instance be ready to conclude that Merfolk beats UWT if a complete scrub picked up UWT and got destroyed a few hundred times, or if someone played it with a budget version?

If you disregard hidden variables, you can for instance show that the amount of polio cases is tied to the amount of ice being sold - which obviously isn't the case. Correlation does not imply causalation; in this case it's that polio triggers more frequently during summer.

In the same spirit, you need to examine your data more throughly - does pilot matter? does deck version? I think it's fairly safe to assume, yes. They matter. Your test needs to reflect this.

yeah, mtg is not too complicated to do statistics on. <snip> It would make for a fun project though, applying what we learn in stat to mtg. I mean really applying it.

You can easily use statistic on Magic, you just need to examine more interactions than just amount of wins. That is of course, if you wish to reduce your model to point at "the best deck".

Compare with quantum mechanics.

Quantum mechanics is just Bayesian statistics in disguise. ;-)

[Resist_Temptation]

I was a merfolk player for a long while, but I just cannot accept the fact that the zoo/goblins matchup is awfullll. After seeing your videos, and the consistency the deck actually has I was wondering if you would have and suggestions as I havea monthly legacy event here and I want to run this but I am 1 tundra short and 1 knight of the white orchid short. Are there any other cards that would fill that? maybe just a cantrip?

[pi4meterftw]

I was a merfolk player for a long while, but I just cannot accept the fact that the zoo/goblins matchup is awfullll. After seeing your videos, and the consistency the deck actually has I was wondering if you would have and suggestions as I havea monthly legacy event here and I want to run this but I am 1 tundra short and 1 knight of the white orchid short. Are there any other cards that would fill that? maybe just a cantrip?

Knight of the white orchid is not in the list. Rather, it's a replaccement for sphinx if you haven't obtained it yet. Are you saying you also don't have sphinx? I dunno, you could try to run court hussar or something.

[Resist_Temptation]

I can probably get my hands on one of the two. They are not very expensive. I was thinking they were going for quite a bit more.

So I ended up buying the rest of the cards..well not the tundra, and was curious about a few things. I am replacing the tundra for now and at first glance I thought I should replace it with an island but the deck has quite heavy W, and that led to me play a third basic plains, but then I ended up missing blue in some test games. I have hallowed fountains, but the number of times I have to replay my lands made me shy away from shocks. How should I replace this one. =P

What is your boarding strategy against survival. (I guess BGR survival and bant survival) Do you take the same gameplan as against ichorid seeing as they both use the graveyard for quite a bit of their engine? ...Well for bant survival I would not board in Forge Tenders..but we play massive grave hate against that archetype correct?