MTGO Vintage Challenge compiled results

Howdy Vintage Community!

A while back I took over gathering results from the weekly MTGO Vintage Challenge and compiling the data from the fantastic Matt Murray. It's been a long time coming but I have a rough draft of the data compilation. It includes:

  1. Tab 'MTGO Results': All MTGO Vintage Challenge Results starting in March 2018 (I did miss a handful of weeks over the last year)
  2. Tab 'Decklists': Decklists from all top 32 competitors
  3. Tab 'Deck Breakdown': Top 32 players and deck archetypes as classified by MTGGoldfish
  4. Tab 'MTGGoldfish Archetype Breakdown': An archetype vs. archetype breakdown that can be filtered over a data range
  5. Tab 'MTGGoldfish Archetype Analysis': A specific archetype breakdown that gives it's win % over other archetypes per week
  6. Tab 'MTGGoldfish Metagame Breakdown': A breakdown of archetype % per week.
  7. Tab 'Card Usage': Full list of cards and occurrence over the full history in the sheet

My intention is to keep this updated with no more than a 1-2 week lag time to keep this as real time as I can.

Feel free to consume as you see fit:
https://docs.google.com/spreadsheets/d/1bywx8Uu5zuPx5jQx7hArmD6bummMOv1tlYSpIt4YGnk/edit?usp=sharing

Some things I'm hoping to add in the future, but feel free to make suggestions and I'll try to keep up!

  1. Data filters for the card breakdowns
  2. Better deck classification algorithm to get archetype and sub archetype granularity in the comparisons - Completed
  3. Some variety of tagging system seen in previous spreadsheets used to compare deck performance with certain characteristics. (taxing effects, null rod, FOW, etc.) - Completed (Thanks @diophan!)

UPDATE: The updated archetype classifications and tagging system is now active. Tabs have been added for the following:

  1. Archetype Breakdown: Breakdown of win% vs other archetypes
  2. Archetype Breakdown - Delta: Same as above but with two data ranges to run the delta for two time periods.
  3. Archetype Breakdown Detail: Details per archetype of match win %; mirror match wins; field win %; sweep %; swept %; game 3 win %.
  4. Archetype Breakdown Detail - Delta: Same as above but with two data ranges to run the delta for two time periods.
  5. Tag Breakdown: Same as Archetype breakdown except for tags: win% vs other tags
  6. Pilot Breakdown Detail: For those of you interested in your historical data.

Thanks!
-Keith

last edited by k0dy

Thanks for taking up the mantle of this!

I noticed for your archetype and tags you have a comment that this "needs an algorithm". When I did this with Matt I indeed used a python script to parse WOTC's webpage and made a determination of tags and archetypes based on the presence of various cards. I can send it to you if you'd like.

That would be excellent, I'd appreciate it greatly! That's exactly what I'm looking to build as my previous method of classification was based on spreadsheet formulas and very painstaking to update. I'll be able to build it into the Google Scripts backing the spreadsheet so it can update whenever new results are imported.

Thanks!
-Keith

Is there a tab for match win percentage? If so, I don’t see it.

I've been wondering that myself for quite a while, so bug survival belongs to the fish archetype?

@smmenen The 'MTGGoldfish Archetype Breakdown' tab has the overall match win % for each archetype, as well as the match win % against known decks. (Followed by the match win % against each other archetype).

@diophan Thanks again for the scripts! I used the classifications and built the archetype and tagging into the spreadsheet formulas. For anyone out there up to debating classifications of different archetypes I am open to suggestions for improving the current setup.

The current systems uses the presence of certain cards to generate a score for each deck, and the highest scoring archetype wins.
If we want to take this further a few ideas I have are:

  1. Adding a weighting factor to each card in a given deck archetype. This would allow certain cards to weight towards the score higher than others (things like Oath of Druids and Survival would be prime candidates for this)
  2. Either adding a sub-archetype that is more specific than the current overall archetypes, or expanding the current archetypes.

The tagging system works at the moment, but doesn't have an associated win % breakdown, but that will be coming in the near future.

This is an absolutely incredible repository of data and information. Thank you so much for putting this together and maintaining it.

A few cool things:

  • The pilot tab has every player who has played in a Challenge in the last year or so.

For players who have more than 30 matches, here are some notable results:

VSL players:

  • @The-Atog-Lord: 62.79 match win %
  • @brianpk80: 61.83 match win %
  • Ecobaronen (Andreas Peterson, easily one of the best players in Vintage right now): 52.17 match win %
  • @Montolio: 65.43 match win %
  • Smmenen (myself) 64.52 match win %
  • Paul Reitzl (Littledarwin): 54.55 match win %

Other notable players:

  • @IamActuallyLvL1 (Justin Gennari): 65.12 match win %
  • @diophan: 63.33 match win %
  • @Thiim 56.52 match win %:
  • Lampalot: 63.33 match win %
  • @ChubbyRain: 70.59 match win %
  • Jazza: 55.93 match win %

That's pretty amazing.

What that means is that, if you are really good at Vintage, your win percentage is probably in the 60s. Or maybe you might just be Chubbyrain, good πŸ˜‰

last edited by Smmenen

@smmenen This is going to make it harder to convince people that my bad decks are actually good or that the bad cards I run are actually playable. 😞

Wow this is an incredible compilation of data. Thank you for putting this together!

Also thank you @Smmenen for tagging me or I would have missed this!

I'll have to dig into this once I get home and can utilize my monitor setup.

For years, I've said that my apprentice and secretary Matthew Murray is the best Magic player I know and alas, my assessment is vindicated. πŸ™‚

Howdy again, thanks for the kind words @Smmenen and everyone who has taken the time to take a peak. I wanted to make a few notes here about changes / updates to the sheets.

  1. While the sheets are locked, it is not fully read only. Any sheet containing date ranges at the top has those cells unlocked so you can update the views accordingly.
  2. Almost all of the formulas are dynamic enough to update whenever I push new data to the spreadsheet, which should now be occurring weekly. (Saturday night for Vintage Challenge data, Sunday /Monday night for published deck lists).
  3. Feel free to make suggestions / ask questions either in this thread or directly in the comments area of the spreadsheet, I am always open to adding features or views that others feel are useful, and willing to help out with any questions anyone has.

Thanks!
-Keith

Thanks! This is really useful, so I'm sure I'll have more questions in the future πŸ™‚

My main question though is just to clarify, that the overall match win % for archetypes is "Known OP Match Win %," correct?

Also, for your metagame breakdowns, are these complete breakdowns or just the top 32 decklists?

@smmenen The 'Overall Match Win %' is the total win% for that archetype over all matches (including those where the opponents deck is not known). The 'Known Match win %' is the win % against top 32 decks, which are the only ones we know.

To be very specific on the data included:

  1. For Archetypes/Decks: We only have the top 32 decks from each challenge, and while we have records of every match for those archetypes ('Overall Match' numbers), we only know the opponent archetype if the match was against another top 32 deck ('Known Match' numbers).
  2. For Players: Because we don't include archetype in these numbers, all matches played by that player that we have record of are included, even if it is using or against an unrecorded deck (non top 32 decks).
  3. Byproducts: I chose to include the overall win% and known match win % in all archetype numbers, as they are both technically skewed in opposite directions.
  • Overall win % (against all decks) is skewed upward because it only contains results for when the archetype top 32'd a tournament.
  • Known match win % (against top 32 opponents) is skewed downward because it is against only the stronger performing opponents.
  • The 'Unknown' archetype represents all of the results for what ends up being 33rd place and below, showing the overall win% for those decks, and the win% against known decks.
  1. Exact data collected:
  • Full match lists with player, opponent, and overall game results from every vintage challenge I was able to pull (I did miss 8 weeks or so over the last 11 months)
  • Full deck lists from every Top 32 Vintage Challenge since 1/1/2018.

While I recognize this does taint the results to some extent, the relative numbers over enough time I think still give a good picture of each archetype's strengths and weaknesses.

A final note: If people think it would catch on, the solution to the missing data problem would likely be the creation of a form allowing players to submit their deck lists (or even just the general archetype) and pushing it to the spreadsheet. I'm hesitant to put the time into it if it's not going to be used, but would still very much like an avenue to complete the missing data.

Links:

Note: I am actively working on a clean solution to the data filtering problem. (If you try to filter the data everything turns to mush because of formula's) The best solution might just be to download the raw data and work from there. To do this:
Click 'File' -> 'Download As' -> 'Tab separated values (.txv)' and then you can either open it in your favorite viewer/editor, or upload/paste into your own Google Sheet.

@k0dy said in MTGO Vintage Challenge compiled results:

I am actively working on a clean solution to the data filtering problem.

You can use "filter views". They change the way we see the data but not their true positions so no problems with formula.

@k0dy Hey, has this been updated lately?

@smmenen Kody said on stream that the new update to the interface broke the data collection. It's a shame as it was a very good resource.

Can the coding be updated so that the latest results can compile?

  • 21
    Posts
  • 3524
    Views