sierkovitz.bsky.social
Data analysis in Magic: The Gathering; Magic Numbers podcast; Limited Resources regular guest. Interested in coaching - ping me a message.
linktr.ee/sierkovitz for socials
507 posts
1,703 followers
908 following
Regular Contributor
Active Commenter
comment in response to
post
twitch.tv/sierkovitz
comment in response to
post
Both changes in frequency and innovation are key mechanisms of self-correction and to convert successfully from early format to mid format drafting - you better keep your hand on the pulse.
And who knows - maybe try innovating yourself? Surely one every 5 drafts can be done for science?
7/7
comment in response to
post
Part of the self-correction is based on frequency of drafting but part will be based on innovation. Some color combinations are easier to figure out early, some take time. If you base your understanding of the format on the early data you risk lagging behind.
6/7
comment in response to
post
Is the format fully corrected? Not yet and it never really does. Orzhov is still underdrafted - I think because it is misunderstood. @samuelhblack.bsky.social - maybe a good topic to cover? Izzet is still overdrafted because - again my hypothesis - the archetype is appealing.
5/7
comment in response to
post
At the same time the win rates got flatter - standard deviation decreased. Top win rate pairs are drafted more, so the decks become weaker and win rate lower.
Underdrafted color pairs get access to good cards more predictably, and their win rates increase.
4/7
comment in response to
post
Play rate standard deviation increases over time. So there are larger disparities in play rate between week 1 and week 2. Players realised what the good and bad pairs are and started drafting more of the good ones and less of the bad ones. And you can clearly see it on the graph below.
3/7
comment in response to
post
This graph shows standard deviation of play and win rate of the 10 color pairs in DFT. This is a measure of dispersion of data: the smaller it is, the closer all the values are together. What does it mean?
2/7
comment in response to
post
Clearly you have not been introduced to Polish cinematography
comment in response to
post
comment in response to
post
Also - who will watch the watchmen? There needs to be a second, more secret group
comment in response to
post
You definitely put Sneaky in Sneaky Snacker...
comment in response to
post
If you are looking for microbial inspiration in solving the problem - here is the link to the article in question:
www.sciencedirect.com/science/arti...
Other experiments showed how bacteria deal with cheating in efficient ways. But the moment policing stops, cheaters are always on the rise.
9/9
comment in response to
post
Because if small accidental errors are given a pass - even if they are accidental, they will inspire others to abuse that in a less accidental way. Surely an occasional game loss due to sloppiness is not a major cost for having a fairer competition at a high level?
8/9
comment in response to
post
We see judges as people who assure games are played according to the rules, but their role should be as people who keep the community free of cheats by making sure cheat-like behaviours are rooted out of the system.
And I use the term cheat-like behaviours very consciously.
7/9
comment in response to
post
Another key part: "For punishment to occur, at least a portion of cooperators must possess specialized cheater-inhibition or killer traits".
Does this not describe judges? Judges are uniquely equipped to penalise cheat-like behaviours. And they imho use that power more frequently. 6/9
comment in response to
post
If cheat-like behaviours are free - they will spread. Also because over time, if cheating is beneficial, we will see more and more cheats in competition, because they will be, on average, slightly more successful. Only way to stop it is to make sure they are not.
5/9
comment in response to
post
Here a fragment from a classic Travisano & Velicer paper on microbial cheating from 2004, looking at the strategy of punishing the cheats in microbial communities.
The key part: "Punishment (...) involves more than simply the absence of benefits that are conferred by social interactions."
4/9
comment in response to
post
In ecology, such behaviours are called cheating. And, if not addressed, lead to bad functioning of the whole community, as over time cheaters spread which hampers community functioning.
But bacteria do address cheating behaviours. And in pretty strong terms.
3/9
comment in response to
post
Cheating behaviour is common in nature. Microbes do it all the time. Some mutants in the microbial community will not contribute to a common good (for example detoxification of environment) because it costs energy. And yet they will try to benefit from others doing it.
2/9
comment in response to
post
Nesting Bot has a natural replacement level ;) So does Basri - maybe that's why I trophied :P
comment in response to
post
www.17lands.com/draft/fa7e2f...
Some really tough picks in the first few packs. In hindsight I should of course pick the wrappings. But how would you navigate it yourselves?
I guess all's well that ends well..
comment in response to
post
Yeah - I sat this one out, I am more interested in draftable product. But I did get my boxes of DSK on Friday and it felt great! Now to make 8 people appear in the same place at the same time...
comment in response to
post
I love historic reenactment like the other person - just recently went to jousting tournament in the local castle. It’s the period they chose I have problem with ;)
comment in response to
post
You can follow me on X too - as I cut all data content - nothing enlightening there, just shitposting, so all but guaranteed you won't get smarter through that.
Food might appear tho ;)
comment in response to
post
You won just by being in this noble company... GL over the weekend!
comment in response to
post
Good luck to all of you and have fun in Chicago - while you are there, find a good spot to sample Polish cuisine. In the end it is the 2nd biggest Polish city. And Polish food is really good for cold weather.
*doesn't apply to Boros
15/15
comment in response to
post
Of course finding an open G lane is still the best, but don't despair if you don't.
WB can be great. UR and BR decks can compete with G ones. Open artifact synergy lane is amazing. Look for those options and who knows, maybe it will compensate for you horrible standard deck choice!
14/15
comment in response to
post
Generally - it seems that at high level all is possible*.
Yes - Dimir decks win less. But Dimir with Haunt the network already has 63% WR - more that generic GB or UG deck. Showing you how tightly packed are those decks. Finding open lane is better than trying to force Green.
13/15
comment in response to
post
One outlier card in the set is Marshal's Pathcruiser. 3 mana is just such a great deal, and in a format where late game is important, a land is a neat thing to have. And if you can crew it to trade it off with something big, while preserving your creatures - you won big time.
12/15
comment in response to
post
U is the tricky one. Best common is Bounce Off. Then we have Flood the Engine, Engineer and Thopterist which are solid+, but lots of U commons are just not that great.
U does exist mainly on uncommon level with some absolute stars (Refueler, Stock Up, Skyray, blue shell and Transit Mage).
11/15
comment in response to
post
Red looks the weakest. But that opens up some potential. If PT players did their homework, they might avoid it. Gunner is by far the best R card at common, but Diver, Strike and Burner Rocket have solid numbers in top player group - if you can get those easily, you might do well.
10/15
comment in response to
post
B has a slew of solid commons - removal, Wickerfolk, Pactdoll, Rat and Max Overdrive, which makes a great impression of Run Over (dare I say Throw from the Saddle?).
All B pairs look very playable, albeit with some caveats (you want to have Haunt the Network to start drafting UB).
9/15
comment in response to
post
G is still great, but the best cards look slightly different. Scurryfoot is the common to play, while Ketradon is doing worse - in higher tiers people come prepared for your 6/6 dumb guy, so you can't rely on it as much as in lower tiers. Part of that prep - Bestow Greatness has great stats.
8/15
comment in response to
post
And W has several outstanding signposts - primarily Beastrider and Ascendant (low key bombs), but also Pathmaker has solid stats - particularly if it can find a win con, like the Riverchurn Monument and attack opponents on two fronts.
White is not great but don't ignore it.
7/15
comment in response to
post
And you can expect W being a bit more open while people are fighting over G - so you can pick key W cards.
W cards like Gloryheath Lynx, Sundial or Glider can present early clock and with support of Ride's End and Gallantry - can win the race. Aggro is not all dead in DFT.
6/15
comment in response to
post
Some key card differences between generic and top player data?
Let's start with some cards that do better in hands of top players.
First - Ride's End. This is the top common for the top players - 63.1% GiH WR. W is not the strongest color, but that card allows W decks fight G efficiently.
5/15
comment in response to
post
You can also see in the graph that several color pairs have a relatively low penalty for splashing, meaning that adding well thought-through cards compensates some of the mana base awkwardness. In the grindy PT drafts that may be a key difference, and knowing which pairs splash better is key.
4/15
comment in response to
post
And normally, those players have smaller differences between top and low win rate color pairs. Same is true for DFT. Only 1%p separates top (GU / GB) and 7th best color pair (RB).
Even UB and WU are just 2%p. behind - a difference but an acceptable one. Boros does lag behind.
3/15
comment in response to
post
First - the big picture. Average 17Lands player is above average, but PT players are a different animal altogether, so generic data is not the best way of capturing them. That's why I use top players data - those are usually >60% win rate players, mostly in mythic.
2/15
comment in response to
post
Does it require a loop when it connects for 10K?