Off Topic: Know thy Analysis



Since last week, I have been testing my pet deck Greenimator through the gauntlet. I replaced Empyrial Angel with Iona, Shield of Emeria. The results were a little weird against Red Deck Wins (RDW). On the draw, my win percentage was only 35%. From it, I have grown a whole new appreciation for Goblin Guide. The percentages are completely flipped on the play. I won slightly over 65% of the games. The very fact my success depends on the die roll is one of those things that bug the snot out of me. I don’t mind mana screw. I don’t mind drawing the incorrect cards, making an incorrect play, or having my opponent top deck the perfect card. The game has a complete randomness about it. Then again, maybe I just need to make a better deck. I take solace in the fact I still make Jund my lap dog.

Answering One Question

Statistics can provide powerful answers. After punching a bunch of numbers into an Excel spreadsheet, I get a bunch of numbers and probabilities. I can analyze and overanalyze to my hearts content. The problem with statistics is the fact they are just numbers. They only give us numerical answers to questions we ask. As with my win percentages, they are the numbers from my question. Can I beat RDW on a full seven? I let RDW free mulligan. As for Greenimator, I took mulligans as normal. My statistics answer one question. I suppose I could have mulled RDW as normal. My stats would have been better. What good does better results do for me? It may make me feel better. However, I don’t want to go to a tournament to lose every die-roll and never watch my opponent take a mulligan. I suppose my thoughts were of planning for the worst and hoping for the best.

I like statistics, but they annoy me at the same time. I can’t tell ya the number articles I read whose statistical analysis is flawed in some fashion. A while ago, I read an article on H1N1 talking about scientist wondering if a more virulent form was running around in Mexico. The article surmised the possibility of a mutation with the high mortality rates. I rolled my eyes because the writer’s didn’t take into account the impact of the differences in accessibility to healthcare, social economic status, nutrition, pollution and education. Has a new virulent form surfaced? I don’t think so. Just the other day, I read a submission by a group of scientist who vowed they found the bad-driving gene. People with the gene had 40% more accidents and were predominantly of European decent. I wonder if anyone notified NASCAR. The problem I had was they really called it the driving gene. They did conclude the possibility it could have some other ramifications in some slight way to other brain functions. Possibly. It was mostly a gene for driving. Not a gene responsible for some kind of higher motor or eye-hand coordination. It was a driving gene.

If you are taking data for play testing, make sure you are answering the right question. Make for damn sure the published results are what they say they are. I can’t recall how many times I see somebody spouting some homebrew new deck beats every deck in the format 80 percent of the time. It is not realistic. Greenimator beats Jund the majority of the time. On the draw with Jund getting free mulligans, Greenimator beats Jund 50% of the time. On the play with Jund still getting free mulligans, Greenimator runs at 75%. Most of this is because I can land a Khalni Heart Expedition without fear of it getting hit with a Maelstrom Pulse before I blow up the landfall count. I can deal with one Blightning, but not back to back. Those numbers are incorrect in a way. That data was from when I was packing Empyrial Angels instead of Iona, Shield of Emeria. Those are two different questions with the only answer from one of them.

Data Collection

Good data results in good analysis. Bad data gives a bad analysis. When I collect my numbers, I use an Excel Spreadsheet. Why do I collect data this way? One reason is I can enter in the results, save it, and add more data later. I can sit down, test a couple of games and come back to it later. It can be more organized than having a bunch of pieces of paper with scribbles on them. Second, I can let Mr. Computer crunch all the numbers for me. It took a little while setting up a spreadsheet, but it is all done. Cut, paste and fill effects are lifesavers. If I want to look back at the numbers, I can just pop open the file and to reexamine them.

Besides my OCD, I collect data on mulligans, wins/losses, land drops, and cards played. Win percentages are obvious. Next, I don’t ever see anybody collecting data on mulligans. I find this sad because decks don’t mulligan the same. RDW hardly ever has to mulligan. Jund needs to mulligan more often with how complex the land drops can be and the range of the cards. An opening hand of 3 Broodmate Dragons, off-color land, and a Garruk Wildspeaker might want to be shipped back. Greenimator would be stuck in the middle. It is important to know. I don’t ever see anybody collecting data on cards. I have it set-up were I can tally the spells played in each game. It all gets calculated into cards played for winning games and cards played in games lost. People can be subjective about their cards. This way, I can see the games were I played an Iona, Shield of Emeria or Sphinx of the Steel Wind that I won the majority of those games. Sadly, games were I played a Bogardan Hellkite I had a tendency to still lose. It might still make the cut though. It is much more effective against other decks like Bushwhacker. RDW or Jund, not so much. I also keep track of land drops to see, well, if I am making my land drops.

Over Analysis

I got all these numbers, percentages, and I can tout them freely on the Internet. Even though I am proposing good play testing, it can be flawed. Not in a data play testing way, but in a practical way. My play testing gives me a good idea of what works and doesn’t. It shows me I need to tweak my deck if I want to improve my match up against RDW. As I talked about last week, I am testing in a controlled environment. I am not taking into account the 30 slightly different Jund decks. Some of which are running Duress. It isn’t taking into account the numerous rogue decks running around. It will happen. A lot of people out there like trying to make their own mark on the tournament scene. It doesn’t matter on the format. Legacy will still see people bringing their Goblins or Affinity deck even though their decks are no longer viable. Even when Jund, RDW, or Bushwhacker is no longer viable with the release of the new set or dominant Archetype, people will still bring those decks.

Beyond all the play testing, it is very easy to over think the situation. I could spend all day analyzing and analyzing the formats. In the end, I can’t prepare for everything. Just think of the people at the last Pro Tour facing the Punishing Fire and Burn of the Grovewillows. Was the analysis flawed? In a way, it was flawed. Not my point though. A person can’t account for everything all the time every single time. I can go to a tournament and never get pared against RDW. I might get stuck facing White Weenie round after round. For my last big tournament, I tweaked my deck so it had an advantage against Fairies. Know what, I never faced Fairies once. Show up at a tournament and suddenly your ill prepared because everyone is running the new Natural Order and Progenitus or Dark Depths and Vampire Hexmage combo. Swing and a miss.

Intuition

Since my preteens, I have been playing games: lots of games. My grandparents had ambitions of me being a priest. I’m not kidding. They would pick me up from the farm on Saturday to be the server in the nightly mass. After church, we would go to their house, eat and then head over to my great grandma’s house right next door. We played Hand and Foot Canasta. If you have never heard of Canasta, it is a very complicated card game. The games would go quite late. It was my life for many years. I swear to this day that I am the only one of the grand kids to ever hear my great grandma. She never swore. Unless, it was a person playing a game of cards with her. I got good at the game. If I started beating them at the game pretty seriously, my great grandma would start swearing up a storm. She even called me a few names. Not so bad until you account for my age. For me, it was like watching Dr. Jekyll turn into Mr. Hyde. It was my sweet great grandma who gave me presents at Christmas and made me cookies. She was not some sailor calling me obscenities. Once the games were done my great grandma would return, smile and ask if I wanted a treat? My grandma would then tuck me in bed. The following morning I would get up and serve the Sunday morning. In any case, the memory still makes me smile.

Numerous articles try to sell an easy 10 step program on how to be a better Magic player. Reading articles help. Getting an education will enable a career. Unfortunately, it isn’t that simple. I can’t always fully elaborate why I make certain choices in a game of Canasta. I’ve just played the game so much. I don’t know if it is some learned unconscious behavior or what. When I look at my opponent and the board, some part of my brain tells me to discard a King. Why? I don’t know. Sometimes those decisions use logic, but not always. In the same light, I have also played a lot of videogames as a child. The Mega Man series was one of my favorites. Even to this day, I can pick up the new release and beat it by the end of the day. Yeah, they aren’t a particularly difficult game series. Nonetheless, I can play the game and for some reason I get an idea a hidden item can be found in the upper-left corner. Why? I don’t know. Part of me believes it is simply an acquired intuition after playing games for all these years. It is something I don’t believe anyone can learn by just reading articles. It takes experience, practice and time. Even players who have only picked up Magic for a short period of time and found success. With a little digging, I am sure those people have experience playing other games. The acquired intuition from those games simply transferred to Magic.

Analyze This

Enough rambling. If anyone wants, send me an email and I will forward my spreadsheet. I tweaked it this week so all deck names, cards, and calculations are automatic.


Just fill in the deck names and cards.


Fill in game wins, land drops and damage. Life totals calculate for you.


Fill in # played per game.


Automatic calculations.


Again, automatic.

Save it in two different spots just in case the code gets messed up. I am sure if you can figure out Magic, you can figure out a spreadsheet. Just remember, more data is better than less.

Comments

Posts Quoted:
Reply
Clear All Quotes