TheWolf's picture
By: TheWolf, Shane Garvey
Jun 16 2021 12:00pm

You may or may not be aware of a website called If you are: great! Skip the next paragraph. If not, read on!

17Lands is a website that takes the raw data from Magic Arena limited games and presents it in an easy to read format. Magic Arena users install a piece of software called a tracker, that reports back to the website details of the draft, such as what colours were played, what order cards were picked in, etc. 

I thought it would be interesting - now that Strixhaven is getting on - to go through the data and see what we can learn about the format. All of the data presented here is correct as of time of writing (June 10, 2021).

What College Is The Best?

One of the things we can look at in the data is which colour pairs are performing the best. 17lands breaks the data down into seperate entries for Quick Draft, Premier Draft, and Traditional Draft, as well as Sealed. Let's take a look at what we can learn.

A quick note: it's useful to know what the average winrates of 17lands users is before we begin. As you will see, the average winrate is higher than what you would think (50%); this means that people who have the tracker installed tend to be better than average players. Keep this in mind. The average winrate across the formats is:

  • Quick Draft 55.5%
  • Premier Draft 54.8%
  • Traditional Draft 60.1%
  • Sealed 54.9%

In every single format, there is one colour pair that wins at a higher rate than the others: Silverquill. In Quick Draft, it has a winrate of 60.8%: 5.3% above the average of all 17lands users. In Premier Draft this is 58.8% (4% higher). In Traditional Draft, which will be of more interest to paper drafters and MTGO users, that winrate goes up to 63.6% (3.5% higher than average). Finally for Sealed, it is 58.6% (3.7% higher than average). This makes it pretty clear that Silverquill is the colour pair to be in.

There is also a clear second best colour pair: Quandrix. The Quick Draft winrate is 56.2%; Premier Draft is 56.3%, Traditional Draft is 62.3%. The only place this is not true is Sealed, where it drops to 3rd behind Lorehold, with 54.6% (0.3% below average).

These winrates only look at 'pure' two-colour decks, without factoring in splashes, though splashing doesn't change the position of the decks in the pecking order.

Should You Splash?

Speaking of splashing, we can look at the data to see if splashing is worth our while. Here is the average winrates of pure two colour decks in the various formats:

  • Quick Draft 57.0%
  • Premier Draft 55.9%
  • Tradtional Draft 61.0%
  • Sealed 55.0%

And here is the winrates for a two colour deck with a splash:

  • Quick Draft 54.9%
  • Premier Draft 54.2%
  • Traditional Draft 59.6%
  • Sealed 53.9%

By splashing, you are shaving 1-2% off your winrate, which isn't heaps. Although the lower winrate, this is not to say that splashing is never correct. 

Is Any Colour Pair Worse Than The Others?

Having established Silverquill as the strongest colour pair and Quandrix second, is there one colour pair that is clearly worse than the others? This is what the data tells us for each colour pair, and each format:

  • Quick Draft: Witherbloom 55%, Lorehold 53.8%, Prismari 52.5%
  • Premier Draft: Lorehold 54.3%, Witherbloom 54.2%, Prismari 54.1%
  • Traditonal Draft: Lorehold 59.5%. Witherbloom 59.3%. Prismari 58.7%
  • Sealed: Lorehold 55.4%. Witherbloom 54.4%, Prismari 51.3%

In Quick Draft and Sealed, it seems that Prismari is clearly worse than the other colours. However this difference evaporates in Premier Draft and Traditional Draft, which are the two 'human drafts' on Arena. The take away here is that if you are drafting in paper or on MTGO, there is probably no clear worst colour pair. 

Play or Draw?

Most people agree it is usually better to be on the play. With the 17lands data, we can tell whether this is true or not.

Unlike the winrates above - where the average is actually higher than 50% - the average for this data is, indeed, 50%. The data tells us that in Strixhaven Quick Draft, Premier Draft, and Sealed, the player going first wins a touch over 51% of the time. In Traditional Draft this is higher, with the player going first winning just under 52% of the time. Therefore, we can agree that in most cases, playing first is correct.

Card Data

With 17lands, we can also look at data for individual cards. This is good information to have, but you also need to be careful analysing it so that you don't draw the wrong conclusions. So let's take a look at some interesting card facts. Note that for this data set, I am going to be using Traditional Draft as a basis, as this has broader appeal for paper and MTGO users. This data is drawn from over 200,000 games played on Arena.

What Cards Are We First Picking?

For this data, we are going to look at the cards that have an 'Average Taken At' rate of less than 1.5, means that more often than not, when it appears, it is taken as the first pick. Here is the list, in order:

It's no surprise these are all mythics (and one rare). So, what about commons? This data is probably more important. If you see these cards wheeling later than their Average Taken At pick, it may be a sign that the colour is open. So, here are the top 7 highly picked commons, with their Average Taken At score:

Also, of note, is the Campus lands. On average, every single one of these goes later than 8th pick, meaning there is a good chance you will wheel them. (This is even more true in Quick Draft, for reference).

The 'Best' Cards

The data also includes a 'Win Rate In Hand' field, which is a way of telling us how many games are won when the card is in our opening hand or we draw it at some stage during the game. Note that this doesn't necessarily mean the card was played, but it is a good way to tell us what the bombs of the format may be. Here are the top 10 cards and their Win Rate in Hand percentage:

How about the commons? The first four are all Lessons: Expanded Anatomy, Fractal Summoning, Inkling Summoning, Elemental Summoning. This makes sense; these cards are all cards that you can get when you need them. The non-Lesson cards are:

The Professor and Hunt for Specimens surprise me a bit here, though the other four I can all understand. The list is also very similar for Sealed.


I hope you enjoyed this look at the raw data for Strixhaven. You can find more data over at