Sunday, September 7, 2014

The Pauper Gauntlet and some thoughts from David Shaffer

Pauper Gauntlet submissions are still open. Please submit a deck in the comments to this post:

Deck submissions will close on September 10th when the Magic Online downtime starts.

Please include a web link to your deck as well as a sideboarding plan. Its all in the article.

At the time I write this we have 57 decks ready to run the Gauntlet but I could deal with a few more original brews that can handle a turn 1 Delver or Mogg Conscripts.

Some thoughts from David Shaffer

I emailed David about using his old article and to see if he had any new ideas for brewers now that daily events are back, and this is what he said...

I'd say a piece of information that isn't ever discussed really is the illusion of testing. But it happens to all of us. The concept is that some people will play a deck and maybe they are bad with it and they are playing against good players. So they lose matchups that the deck should normally win. Then you have other people that are good and play against bad pilots, giving them false positives when they win. 

Since I left Pauper, I've periodically checked in on people playing various decks. I find a common theme is that people say something like Deck X is 50/50 vs. Deck Y. Or Deck Z is 40/60 against Deck A and so on and so forth. I want to caution you against using this data improperly because you want to avoid consuming bad information as much as possible. If you think your deck is good against MonoB because of what your buddy Doug said. But he came to that conclusion because he played against awful players, then when you actually go up against good MonoB opponents you'll lose. This can happen even if Doug isn't there telling you about the match-up. It can happen if you test against the wrong people. As I'm sure you can see, this can be highly problematic. 

I've obviously done this too, and have fallen victim to these same data issues in the past. With AZkitty, I thought it was good against Delver, because my initial experience with the deck in (real) Daily Events lead me to conclude that. I initially entered into Daily Events with the deck and did well against the occasional Delver pilots I ran into. Mostly, because, I believe, that the Delver players on the whole weren't that good. I knew this, but when I beat Mezzel, who I regard as one of the best Delver players, I came to the conclusion that Delver was a good match-up. When we switched to Premiere Events (PEs) I ran, not only into a lot of Delver, but also a lot of good Delver pilots. One of the big issues with PEs was it became apparent that Delver was the best deck when piloted correctly. I got ample opportunity to run my deck into Delver lists over and over again, and later concluded that AZkitty is actually slightly unfavored against a capable Delver player. In my hands the deck was still, however, good against unskilled Delver players and got wins against good players via variance. Throughout my experience in Pauper, people who would also play AZkitty would tell me that they couldn't ever beat Delver, but others would tell me they couldn't lose against it.

From all of this I've gathered that AZkitty is simultaneously good against Delver and bad against Delver, but it really depends on who the pilots are. As most match-ups are in the current Pauper meta-game, it isn't a black and white affair. 

All this shows not only the importance of proper testing, but also the need to know the matchups for yourself and to estimate your opponent's expected skill level before entering a tournament with your deck. 

1. Proper Testing 

You need to test a lot and against a wide variety of players. I think a lot has been said about the quantity of testing. You need to do a lot of it. But not as much has been said about the quality of testing when working on a new deck. If your friend Steve is better than you, and you play Steve in 100 matches you may conclude every deck you play is worse than the deck Steve is playing. From an improvement perspective, I think it is a good thing to be playing Steve because you will get better at playing. You want to play people who are better than you in order to improve.

But if your goal is to work on making your deck design better then it may be the case that you want to play people of varying skill-sets. For instance, if Steve is a pro,and pro caliber players make up only 5% of the field then you should probably play other worse people too. If Steve wallops on you, it may give you a false impression of how your deck would perform in an upcoming tournament. His skill-level would overshadow your deck's performance. So you should be testing your deck against players you are expecting to play against. 

2. Know the Matchups for Yourself 

When looking at what everyone else is going to bring to the tournament, think about how the matchups shake out in your head. Maybe after you've tested your AZkitty list against Delver 100 times you find, regardless of the opponent's skill level, you still lose a lot. Well if you expect a lot of Delver in a tournament maybe it is time for a deck change. The opposite can also be quite true. So keep these both in mind. 

My biggest frustration is not when people say stuff like MonoB always beats Affinity, but it is when people use that information improperly. When someone says "MonoB always beats Affinity" what they are really saying is "When I've played MonoB I always beat Affinity" or "when I play Affinity I always lose to MonoB" or maybe it is simply "When I see these two decks on paper, MonoB looks like it should always beat Affinity." But none of those statements are true in and of themselves. What you need to do is figure out if it is true for you. If you want to play Affinity then see if MonoB is unbeatable. But don't just take someone else's opinion as the truth. Instead, use their opinions to corroborate your findings. Maybe their statements can help you determine more rapidly whether MonoB is a bad matchup, like maybe you only need to test against it 90 times instead of a 100, but definitely don't take their opinion as gospel. 

3. Estimate your Opponent's Expected Skill Level 

Remember Steve who I was talking about above? Remember how it doesn't make sense to test solely against Steve, because he is a Pro? Well now imagine you're about to go to the Pro-Tour. Suddenly it makes sense to test solely against Steve. Now, he's actually the level that you would expect to play against, so him walloping you may be more valuable for you than it would be if you were testing for one of these new Weekend Daily Events (WDEs). 

As you can see, the last thing you need to do is to determine how good you think your opponents are going to be in the tournament you register in. If you want to bring a deck that you can't beat a good Delver pilot with, we know you definitely don't take it to a PE. But PEs are no longer an option. So what are we expecting out of a WDE? Could we bring a deck like AZkitty there? Quite frankly, the average player in a WDE is worse. Because the payouts and tournament structure are more conducive to the Average Joe player, more Average Joe players will, and do, enter WDEs. This is why DEs are so much more conducive to innovation and deck design. As such, this should impact our analysis of our deck. 

But exactly how much worse is the player base in a WDE than a PE? Well, we certainly aren't testing against the Steves of the world. To my knowledge, only a handful of Pros and Pro caliber players play Pauper, and of those most just dabble in the format. There are some "grinders" that you will run into and they are on average a little below the Pro caliber player, but they are still good. While I've never really tracked this, back in February when we had PEs firing semi-regularly, I'd guess the number of good to great players was close to 50-60% of the tournament base. I'd also guess that right before PEs stopped firing the percentage of great players in the queues was probably approaching 80%. Because WDE's don't have a prize structure that scales with how good you do, I'd say that you can safely expect to see 30-40% of the queue filled with strong to great players. Again, I haven't really tracked this and I haven't logged on to Modo since June, but my gut says this number is about what you can expect based on what I played against in January - October 2013's DEs and January - March 2014's PEs. 

So after compiling all that semi-pseudo data above, I'd put the WDE skill level at just above tournament practice room (TPR) quality. I think it makes sense for everyone to be playing in WDEs with any brew they feel is competitive after some moderate testing.  If you have a brew, WDEs are actually a very safe place to test. What I would do is try to determine how good your deck is by first starting in the tournament practice room. If you're starting to see the success numbers you want , then you might want to jam a few 2-mans or 8-mans, but arguably the next logical step in testing is to go straight from the TPR to WDEs. You should expect a slight increase in skill level. I'd guess that at least one of your matches each tournament will be against someone good. But even if you don't think you're a great player, that still gives you 3 other matches you can win. Plus, don't forget that you still can beat the good player with variance. Then, even if you do scrub out of that WDE, with $3.00 pack prices, you still have 2 more shots at winning before you're out of a 3-1 full payback range. So in short, your WDE opponents will probably be average caliber players. 

No comments:

Post a Comment