mattingly23's picture
By: mattingly23, Mike
Sep 07 2015 12:00pm
1
Login to post comments
2911 views


Introduction

In my previous article, I gathered statistics for a Pauper Daily Event based around the most cast spells overall as well as in the first two turns of the game.  This article will look at another daily event and provide similar analysis, with the addition of some new statistics.  The data for this article comes from the Pauper Daily 2015-09-05 event. In total, this covers 102 games consisting of 30 individual players.

One thing to note is in regards to the sample size of the data we're looking at.  So far this is only the second event from which I've gathered statistics.  Over time as I collect data from more events, I can compile them to get a less variable view of the format.  While an individual event may present more variation, it can still give us some window into format.

Feedback from Previous Article

In regards to feedback from the last article, I can say one of the lacking details in these statistics is that I'm not providing any information on the archetypes being played.  Given what we know about the format, we can make educated guesses on what decks many of the cards appeared in.  However, for those versatile cards being played in different types of decks, looking at data based purely on cards doesn't bring the best clarity.  I have some ideas on how to get more views into the data based on archetypes, so look for that in future articles.

Many of the comments I received in the Pauper subreddit post of the article pertained to Cloud of Faeries and the Familiar Combo.  Since these have been hot topics recently this was not a surprise.  As we look at the stats I'll provide some commentary specifically on these topics.

Most Cast Spells - Overall

The table below consists of the top 20 most cast spells.  The first three columns of data (Total Cast, Cast in Winning Games, Percentage Cast in Winning Games) correspond to the same columns I presented in my previous article.  The next three (Total Games Cast In, Winning Games Cast In, Win Percentage of Games Cast In) are new.  These new columns were added to show another view into the data, one that is not affected by casting a spell multiple times in a single game.

Card Name Total Cast Cast in Winning Games Percentage Cast in Winning Games Total Games Cast In Winning Games Cast In Win Percentage of Games Cast In
Delver of Secrets 105 64 0.61 49 30 0.61
Preordain 88 46 0.52 54 27 0.5
Cloud of Faeries 72 46 0.64 32 19 0.59
Ghostly Flicker 69 66 0.96 8 6 0.75
Counterspell 62 32 0.52 35 17 0.49
Spellstutter Sprite 54 36 0.67 27 16 0.59
Ponder 52 27 0.52 32 17 0.53
Gurmag Angler 48 28 0.58 22 12 0.55
Spire Golem 47 28 0.6 25 13 0.52
Lightning Bolt 46 28 0.61 34 19 0.56
Snap 44 24 0.55 29 16 0.55
Prophetic Prism 43 37 0.86 13 9 0.69
Mulldrifter 38 29 0.76 18 11 0.61
Abundant Growth 36 25 0.69 16 10 0.63
Kor Skyfisher 35 24 0.69 17 9 0.53
Rift Bolt 34 20 0.59 22 12 0.55
Compulsive Research 31 21 0.68 18 10 0.56
Chain Lightning 30 14 0.47 26 13 0.5
Incinerate 27 16 0.59 20 12 0.6
Capsize 26 15 0.58 4 2 0.5

Full csv on Pastebin.com

This time Delver of Secrets took the top spot as the most cast spell, with Ghostly Flicker moving from first to fourth.  In the previous event I covered, two players went 3-0 with the Familiar Combo, while in this event only one player (2-1 record) had played the deck (one other deck, an Azorius Tron build, also featured the card).  Looking at the Win Percentage of Games Cast In, players were 6-2 when they cast Ghostly Flicker during a game, taking the highest win percentage of the top 20 cards. A caveat to this noted in the reddit comments on the previous article is that a player is going to be less likely to cast Ghostly Flicker until they're ready to combo.  I'll get back to this card when I discuss "noncast" stats later on.

Most Cast Spells - Turn 1

Here were the most cast spells on Turn 1:

Card Name Total Cast Cast in Winning Games Percentage Cast in Winning Games
Delver of Secrets 26 17 0.65
Preordain 10 4 0.4
Abundant Growth 7 6 0.86
Chain Lightning 7 1 0.14
Ponder 6 3 0.5
Springleaf Drum 5 3 0.6
Chromatic Star 5 2 0.4
Cloudfin Raptor 5 2 0.4
Lightning Bolt 5 1 0.2
Lava Spike 3 2 0.67

Full csv on Pastebin.com

Delver of Secrets returns to top the charts for a second time. Abundant Growth has a strong showing, with players going 6-1 when casting it in the first turn.

Most Cast Spells - Turn 2

Card Name Total Cast Cast in Winning Games Percentage Cast in Winning Games
Cloud of Faeries 15 7 0.47
Rift Bolt 14 8 0.57
Preordain 13 5 0.38
Delver of Secrets 12 7 0.58
Curse of the Pierced Heart 8 4 0.5
Keldon Marauders 8 3 0.38
Ponder 8 2 0.25
Abundant Growth 6 4 0.67
Needle Drop 5 4 0.8
Brainstorm 5 3 0.6

Full csv on Pastebin.com

While Delver of Secrets had been the top card at Turn 2 in my previous article, the suspect some thought would be there, Cloud of Faeries, leads the pack this time.  The casting of Cloud of Faeries on Turn 2 has been in debate as this "free" 1/1 flyer leaves a Delver deck open to follow up with a Counterspell or Spellstutter Sprite, shutting down the opponent early in the game.  In this event, Cloud of Faeries cast on Turn 2 weren't even in winning games 50% of the time.  Though we don't know the Delver/Familiar breakdown, in this event it doesn't seem like a Turn 2 Cloud of Faeries resulted in an outrageous win percentage.  For the event featured in my previous articles, players who cast this spell on Turn 2 went 2-2.

Most "Noncast" Spells

While the data above looked at the when spells are cast, what about looking at when a player has been unable to cast a given spell?  Would we see a strong negative impact when a player was unable to cast particular spells in a game?

One thing to note here - if a player never cast a spell in any game, but they did have it in their deck, that would not be counted here because I would not have it on record that they had played it in their deck.

Card Name Total Games Not Cast Won Games Not Cast Win Percentage of Games Not Cast In
Dispel 34 16 0.47
Ninja of the Deep Hours 30 19 0.63
Electrickery 29 17 0.59
Stormbound Geist 28 14 0.5
Hydroblast 28 13 0.46
Deprive 27 14 0.52
Gush 25 14 0.56
Pyroblast 25 13 0.52
Annul 25 11 0.44
Lightning Bolt 25 11 0.44
Preordain 23 14 0.61
Ponder 20 14 0.7
Counterspell 20 10 0.5
Delver of Secrets 19 9 0.47
Curse of the Pierced Heart 18 9 0.5
Disfigure 17 9 0.53
Lava Spike 17 6 0.35
Cloud of Faeries 17 5 0.29
Exclude 15 9 0.6
Searing Blaze 15 9 0.6

Full csv on pastebin.com

Many sideboard staples are featured in this list since they are going to be brought in only when needed.  Dispel, Electrickery, and Stormbound Geist for example fall into that category.  Ninja of the Deep Hours  is usually brought into the battlefield through its Ninjitsu ability rather than being cast, which would explain the high number of games in which it was not cast.  Its Win Percentage of Games Not Cast In (0.63) is almost double its Win Percentage of Games Cast In (0.33, 3-6).  This seems logical as the Ninjitsu ability is cheaper than casting the card, is most likely on a creature that will deal combat damage to the player so the player gets to draw a card right away, and has the potential to return a valuable creature to cast (like Spellstutter Sprite) to the owner's hand.

Let's catch up with our friends Cloud of Faeries and Ghostly Flicker again.  Players who cast Cloud of Faeries in any game went 5-12 in games in which they did not cast it.  This compares to 19-13 in games when they did cast it.  That's a huge swing in win percentage.  Clearly for Familiar Combo, Cloud of Faeries is essential and certainly if a pilot of the deck was not able to cast it, their win percentage would be hit hard.  In Delver, the benefits are leaving open lands for more spells and assisting Spellstutter Sprite in countering spells as well - great benefits but maybe not the primary driver of the deck.  Since I can't generate the stats for this for specific archetypes, I can't say for certain the breakdown of impact on each deck.

Regarding Ghostly Flicker, players who cast it in any game went 2-6 (0.25) in games in which they did not cast it.  Like Cloud of Faeries, Ghostly Flicker is essential for the Familiar Combo and it's going to be tough to win without casting it.  As mentioned earlier two decks featured the card, one Familiar Combo, and the other an Azorius Tron deck.

Conclusion

As mentioned earlier, a future improvement will be get data on archetypes rather than simply cards.  In addition, I think another interesting addition would be using the data from MTG JSON to use the mana cost and other data there to get some more insights.  Thanks for reading, and please leave your questions or suggestions in the comments!

2 Comments

Data Analysis by ComixWriter at Tue, 09/08/2015 - 13:38
ComixWriter's picture
1

A hallmark of academic research is to find a solution to a posed problem. Data analysis is a key element in this strategy. You've shown us data, but failed to even risk a possible conclusion. You delved (pun intended) into the research for a reason, at least in your mind, but what does the data suggest to you about your initial hypothesis? I applaud your data collection and pretty presentation, but am annoyed that you stopped at the finish line to look around and smell flowers.

Based on your own data collection and analysis, what does your gut - if not mind - tell you about the cards used in the format? We currently fight Delver and MBC in pauper - will that metagame change or needs be changed? How, if so, and with what new conditions may you suggest?

...or are you sticking to a journalist mindset- just report facts without bias? If the latter, I've no reason to read anything further that you write.

Hi ComixWriter - thanks for by mattingly23 at Tue, 09/08/2015 - 19:47
mattingly23's picture

Hi ComixWriter - thanks for taking the time to leave feedback. I'll keep in mind providing more of my own personal commentary and as well as my impetus for collecting the statistics for next time. Even so, I hope others find value reviewing the statistics for themselves, whether out of pure curiosity or so that they might gain some insights of their own (that hopefully they share with the rest of us). If you have any recommended resources I might find useful in the area of Magic statistical analysis, please share!