Peasant Tournament - Final Standings

Discussion in 'Card Hunter General Chat' started by Flaxative, Jan 11, 2014.

  1. TheShadowTitan

    TheShadowTitan #3 in Spring PvP Season

    ItsEtai used 2 elf warriors and 1 dwarf wizard and CT5 used 1 dwarf warrior, 1 human warrior and 1 human priest I think.
     
  2. CT5

    CT5 Guild Leader

    Yup, 1 dwarf warrior, 1 human warrior, 1 human priest, with a mostly useless sideboard =P
     
  3. Flaxative

    Flaxative Party Leader

    Hey guys! I don't have a particularly eloquent essay about the tournament, but I have some stats, a tiny bit of analysis, and some thoughts on the peasant format and the tournament process.

    First up, stats!

    [link] Google Spreadsheet

    The above-linked spreadsheet has the stats of the 13 players who finished in the tournament. I did not include the stats of players who dropped, especially since most of them had played 0 games, and I did not include stats of matches that didn't happen (e.g. Stexe got a win when his opponent didn't show up one round; that is not shown in his match win %). I did include wins and losses against players who later dropped (e.g. ArcadianRook beat Red Mage 2-0, and that is reflected in his stats).

    So this spreadsheet shows us a couple fun things.

    1) Only two elves participated in the tournament, together in a single build. The player running them had the highest win rate in the tournament.
    1A) Said player should be very mad that he missed his third round match, given his win rate.
    2) Triple wizard was the most competitive party composition, though there was quite a bit of variety in deck lists.
    3) Priests were underplayed, and teams including priests did the worst on average. Do priests require more skill to pilot than other classes?
    3A) That said, the player who won the tournament ran a priest, as did one of the players who tied for second place.
    4) Winds of War was probably the single strongest card in the tournament. At all rankings, players who ran a lot of winds had better win percentages than players who ran fewer winds. The one exception to this rule is the aforementioned elf-lover.

    Of course, these stats have more factors than I've quantified. The specifics of the deck list and sideboard list are pretty important, as is player skill. Things like map selection also matter, and apparently going first is a huge disadvantage in Card Hunter (note how every PvP map leans toward Player 2; this was the case with the Winter maps as well, across a sample size of tens of thousands of games). But it's hard to quantify all these things, so I'm going to gloss over them for now. I expect my critics will capitalize on this but whatever. I think the data is still interesting even if it's of limited use. At the very least, it might inform players' decisions as we continue to explore the peasant environment.

    I won't be publishing deck lists at this time just because they're giant walls of text. If anyone is interested in a given build, I can provide it on a case by case basis.

    Implications for Peasant

    I did do a cursory examination of which items were run most. These items tended to be run by both winning and losing teams. Runestone and Bejeweled Shortsword were two items that people were complaining about before the tournament started—basically, since the inception of peasant—and they showed up, along with crowd favorites Double-edged Sword and Perfect Toughness, at all rankings. Not every team ran all the best items, but there are clearly some preferred items. Our data's honestly too limited to show whether Triple Dwarf Wizards—which, no matter its manifestation, relies on trait cycling to achieve above-average consistency—is actually the best build in the format, but it certainly hints at that. Most equipment slots besides the ones occupied by the aforementioned items were wildly varied across decks, though wizards understandably saw a lot of Electroporter Novice and Robe Of Lightness.

    I'm not ready at this time to take any measures against any particular gear. I fully intend to ban problem gear from peasant when it becomes abundantly obvious that doing so is a necessity, but simply throwing out the best items because they are the best seems like the start of a race to the bottom. A lot of gear in Card Hunter is simply bad, and we can't create a format where all gear is on equal footing. I'm not remotely interested in doing so. That said, if wizards continue to dominate the format ("dominate"), I will look into banning small pieces of their arsenal in order to level the playing field. That's just the way it'll have to be.

    In summary: the peasant format is not changing. Yet.

    Reflections on the Tournament

    Here's a little index of what I intend to discuss. (Warning: each is a wall of text in its own right.)
    1) Maps/stages
    2) Match Format + Time control
    3) Sideboarding
    4) Logistics

    1) Maps

    So, I chose 10 maps for players to play on. Each game was on a random stage, though I fudged it so that no match would have the same stage more than once. I thought the maps I picked were a pretty good selection (obviously): some of the more balanced maps from CH's history, and some of the harder/more tactically-involved maps. My understanding of these maps was somewhat flawed—some of the maps heavily favored some builds where I didn't expect them to. On top of that, some players just hated some maps, and complained about the balance flaws of those maps when they didn't actually seem particularly unfair. When matches are Best of 3 and you have a lot of matches, a large rotation seems good, which is why I chose 10 maps. In the end, we got a lot of decently varied games on a lot of maps, and I think that was probably fun for the spectators and hopefully some of the players. In the future, it'd be neat to run tournaments on user-created maps, as was suggested by Tuknir and SLG, but it seems like balance would remain a potential problem. No matter what maps we choose from, tournament directors are going to need to make selections, and when doing so will need to consider all the points raised here.

    As both an organizer and a player my personal preference is toward a large rotation with good variety. So I kind of like the stage-striking rule we implemented in the final round. This rule allowed players to, for a given match, strike one map from the rotation. This allows players to avoid their most-hated maps while maintaining good variance in the stage environment. In the final round, some players struck maps at random (not having a huge preference); some didn't strike any maps (having less of a huge preference); many struck the maps that either they felt were worst for their builds or just didn't like. Based on a single round, I like this rule and how it performed, and think I would use it in future tournaments, regardless of what the map pool is (so long as there're 5+ possible stages, probably).

    2) Match Format + Time Control

    Matches were best of 3, so to win a match a player needed to win 2/3 games. I loved this part of the tournament, and I think a lot of spectators did, too. A player could win a game—perhaps luckily?—and then lose two, which is dramatic. It decreases variance, too: it's possible to lose games off spectacularly terrible draws, and that's more likely to cost you a match in best of 1 than in best of 3. I think going best of 3 is basically the best way to go. Much as Card Hunter is fun, the fact that ranked matches are best of 1 bugs me. It feels like a flaw. Best of 3 enables things that improve the metagame, like sideboarding (the next point), too!

    The big problem with best of 3 is that matches take a while. If each player has 20 minutes on the clock, then you have a maximum game time of 40 minutes, and a maximum match time of 2 hours! While no doubt many games would be played in less time than that, the logistics get a lot harder when you're trying to divvy up 2-hour slots. So I limited time control to 10 minutes/player (maximum match time of 1 hour), and a couple players ended up having time issues. That was sad and I felt bad about it, as someone who likes taking time to think about his plays. But I wasn't sure how to find more time for the players while preserving the benefits of best of 3. Ultimately, a better chess clock, which does something like add 10 seconds every move, would mitigate time issues while keeping matches short, but that requires coding on the part of the devs and isn't within my power. The only real thing I can do to reconcile the needs of best of 3 with longer matches would be to have other people helping me (more on that later).

    In the end, I think the time control was a necessary evil for this event in order to rake in the gains on the best of 3 match format, and I think it was worth it. I'm open to other suggestions regard match format, of course, if anyone else has other ideas.

    3) Sideboarding

    Sideboarding is great. It just is. Our sideboarding rules weren't ideal, though. They should have been, rather than 10 items, capped at 30 cards' worth of items, or possibly a bit less. So yeah, sideboards were a bit too large, but in some matches they created a pretty good dynamic and the existence of some extremely good anti-wizard items might have helped to reign in triple wizard builds' absurd average win rate a bit after the board. It'll be interesting to see how sideboarding evolves in my Card Hunter tournaments, especially with the upcoming new content—more options for purging, punishing card draw, stopping step attacks, and ignoring blocks are all being added to the game, which means more fantastic sideboard cards! I'm looking forward to that :) And I would strongly recommend that anyone else who wants to run a tournament consider including sideboard rules.

    4) Logistics

    The biggest problem with the tournament was definitely on the logistical side of things. Basically I tried to accommodate a lot of my own desires. The various criteria added up to a situation where I needed to coordinate games between wildly different time zones, Australians and Europeans and Americans, and then show up to all these games myself. This was hard with a mere 16-person event; it would be untenable for a follow-up larger event.

    The criteria were as follows:
    A. I wanted the event to be open to ALL players, regardless of this funny thing we call time zones.
    B. I wanted the event to have a referee, because there is no code to check whether players are using the build they registered for an event. Also I wanted a referee to be able to resolve disputes and to keep an eye on players & spectators.​
    C. I wanted the event to be recorded, for there to be video footage available to spectators.​

    A seemed important, though I'm a bit foggy on why. Certainly if I had just said "these are the times we'll have matches," my job would have been a lot easier. But that's neither here nor there.

    B is important mostly due to a lack in support in code. Honestly anyone could have refereed, so this is something I could easily delegate if I had helpers.

    C is a lot harder to ensure. I have far from the best set-up for recording, and I have no idea if my commentary is at all interesting to viewers, so maybe the footage is useless, but it seems cool to have it, and not everyone could do it—you need a mic, a computer capable of running recording software at the same time as Card Hunter, a screen big enough to cast enough of the game, etc.. So while it's easy to delegate refereeship, I can't quite just say "players, here're your pairings, do your games whenever and record them."

    I want to run more tournaments, but the time involved is the biggest issue. Running an event the way I just ran this past one is honestly a job, and while I love Card Hunter I'm not sure I can justify putting in this kind of time without real compensation. Especially not for a larger event, where ostensibly the work would balloon. If anyone has suggestions as to how to reconcile A, B, and C efficiently, I'd love to hear them. Also, since I fully intend to run more events, I'll be looking into ways to efficiently delegate parts of the process, and get other organizers on board. But yeah, the demands create a situation where the work is enormous. I'm very happy with what Jon has given me as a token of thanks for running the event, and I don't mean to sound like I'm complaining—this was a test run, an experiment. And the experiment happens to show that running a 16-person tournament like this one takes more than 20 hours of time over a few weeks. I can only imagine how the time involved would scale up with the number of players.

    Also, I think future tournaments should have entry deposits so that players feel accountable and show up to their games. Players approaching the tournaments too casually to bother turning up to their games probably won't pay the deposit in the first place, and solve a lot of the visible issues with the tournament logistics.

    Anyway, that about sums up my thoughts on the tournament for now. I'd say the event was a success overall, it made a lot of people happy and was new and exciting for the game. We all learned a lot from it, and there will be more events in the future. Thanks everyone who's read this far; hope the wall of text wasn't too boring, and let me know if you have any questions or further topics of discussion related to the event.

    Cheers!
     
    dashv520, Stexe, Lusus121 and 5 others like this.
  4. PaladinGP

    PaladinGP #1 in Spring PvP Season

    Great report.

    How about all players to referee one match per round (usually only doing 1 every 2 rounds) that they're in the tournament? That'd spread the load until there's code for it, which shouldn't be too hard to write given there's already code to "Card Hunter certify" Quest builds.

    Alternatively, you could have self policing, where players can call out their opponent for using an illegal card, but that does distract from play. Also, I think some things might slip past, there: if a player sideboards in, for example, a Shielding Token as a new arcane item, we can see on the Nimbus when it's played that it's from that card, no problem. But when a second Nimbus from a shielding token is played, how can we check the player hasn't sneakily sideboarded in two Shielding Tokens and got one Nimbus from each? Sure, they'll get caught if a third turns up, but it is a potential loophole. I can't see it really being a problem though - the chance of being caught is probably greater than the expected increase to your win percentage.

    Prioritise you not needing to be there for matches: your enthusiasm is important to protect! The more widely a load is distributed - pass it on to the entrants, and penalise those who don't do their bit - the less there is for any individual.
     
    Flaxative and Farbs like this.
  5. Flaxative

    Flaxative Party Leader

    You make some good points, distributing refereeing among players seems a bit problematic though since they could have conflict of interest. And even once we have code to certify the peasant legality of a build, I don't know how it would check that it is not only peasant legal but also the exact build registered for a given event. I guess the API could help with that :) Having an unbiased third party present at every game to ensure that everything is above board seems necessary in any case. Also, recording remains a problem. But yeah! Definitely need to think more about these things, and like I said, I'll be sure to get helpers for the next tournament. Just need to think of how best to delegate :)
     
  6. tuknir

    tuknir #3 in Spring PvP Season

    Interesting, we where discussing this the other day on the chat. I usual prefer going second and imo its usual best, but i guess that depends of type od deck and even the maps so i guess it ends to balance it self

    Anyway great analyses and great work Flax
     
    Flaxative likes this.
  7. Megadestructo

    Megadestructo Shark Card

    Thanks for running this, Flax! This was very cool and I hope we can get some things sorted to make running tournaments (like this one) easier for anyone who wants to try their hand at it, including us!
     
    Farbs and Flaxative like this.
  8. Farbs

    Farbs Blue Manchu Staff Member

    This was a very ambitious tournament, and I'm really glad you managed to make it work.

    I think it'd be interesting to run a tournament geared toward ease of admin, just to compare the two. The easiest tournament setup I can think of is:
    • Single contiguous event, held over a couple of hours one evening
    • Forum thread advertises time in advance, collects registrations
    • Players are expected to be in the lobby until eliminated, may forfeit matches otherwise
    • Matches run in parallel to speed things up, referee(s) on hand in the lobby and may spectate if called, but are not expected to watch each game
    • Double elimination, using a service like challonge to maintain and share the fixings
    • Players use Casual games rather than Custom, to simplify game setup and remove the need for referees to check the boards
    • In-game battle results list and/or filtered API data used by refs to collect/check results
    • Best of 1 rather than best of 3 (optional)
    • No deck limits
    This seems like it'd be a slightly less fantastic tournament, but the effort:reward ratio would be far, far higher.

    Regarding future code support, my thoughts are:
    1. Fixing the timeout issues with the item API requests would make build restrictions possible, though it's still a lot of work to verify them all by hand. Some level of external automation would be possible and helpful there.
    2. Player filter for battle API would make finding and verifying previous results much easier.
    3. Publishing a hash for custom scenarios via the API would allow people to verify that tournament scenarios had not been tampered with.
    4. Gold and/or Pizza gifting could be very useful for collecting entrance fees and awarding prizes.
    5. For extremely large tournaments it might be worth automating the results, either by someone writing their own fixing/result management website, or by someone writing a bridge between our API and the API of a service like challonge.
    I expect #1 and #2 will be high on Tess' priority list for the next API revision, and I'll talk with her about #3. #4 is on our wish list and gaining ground. #5 can be done by anyone!
     
    Lusus121 and Stexe like this.
  9. Stexe

    Stexe #2 in Spring PvP Season

    Yes, I hope for some more tournaments and such, but there are a few issues with making it viable and streamlined.

    The problem with a single event is that it would have to be LONG to really give an accurate representation of who the best players are. Maybe only have it for like 2 hours and then split it up to 3 different events. Another issue is time zone problems (which was a huge issue in the one Flax ran) -- having it all at one time means a lot of people won't be able to participate.

    Having no deck limits means that the people with the highest Elo in ranked currently would most likely just win the tournament. It wouldn't be much different from the way things currently are. Limiting it to Peasant means anyone has a fair chance and is a lot more interesting -- you'll see a lot more diverse builds.

    Really, what needs to be done is have an API that can monitor games and check to see if certain conditions are met (only legal items, map that players are told to use, who wins, etc.) and then give reports based on the game. That way there doesn't need to be a referee and the API can check if everything was legit. Granted, it isn't the same as having someone commentate and record the matches -- but it would be a nice stepping stone. Maybe work on having the ability to save matches and replay them in the client? That way people could both check with API and manually re-watch games for issues.

    Lots of work would needed to be implemented for tournaments to be something that works well... considering big places like League of Legends and such have a huge number of people who do nothing but monitor tournaments it seems tricky for just Card Hunter.

    I might run a Peasant or "Normal" (Common, Uncommon, Rares) tournament in the future, having people simply screenshot the start and end of matches. The problem is checking for legal builds and maps -- which I think should be one of the most important things that API needs to be able to check.
     
  10. Flaxative

    Flaxative Party Leader

    Yeah, replay functionality would actually solve basically all problems. If both players simply meet and play their match, then post the replay, you could referee retroactively and anyone could watch the match at any time. This is also a good feature useful in a bunch of other contexts, so maybe the devs could look into it.
     
  11. Stexe

    Stexe #2 in Spring PvP Season


    Yeah, back when LoL was in beta people were all demanding replays. It took a few years later to convince Riot to get replays in... and technically they still don't have full replay functionality (although on the Public Beta Environment they do). The amount of things that replay adds is huge... such as tournament set ups, watching top matches whenever, analyzing games play-by-play, and more.

    I'd say the ability to save and load replays should be high priority, followed by API for searching item/card use in battles, API for verifying non-tampered maps, and API filtered by player.

    Most LoL tournaments are arranged by the participants and they work out times that both people can play -- it seemed the best when both people talked back and forth to find a time that worked instead of someone saying "you be here at X time." Although that might have a place in there as well.
     
    Flaxative likes this.
  12. Farbs

    Farbs Blue Manchu Staff Member

    Replays do sound fun, and by abusing our network architecture a little they probably wouldn't be too hard to implement, but setting them up to play using rulesets and data from previous builds of the game would be a nightmare. Otherwise we'd have replay files that stop working, or perhaps work fine unless you play particular cards or characters, every time we updated the game. Hmm.

    What does a referee need to do with a replay other than verify that a match took place between particular people, on a particular map, with a particular winner, using particular items? It seems like a simple, functional UI on top of the API would cover all of that, and would be much quicker and easier than watching a replay.
     
    Pengw1n and Sir Veza like this.
  13. Flaxative

    Flaxative Party Leader

    You don't need the replay to referee; it makes other things easy/possible, like spectating or recording footage without being at the match. I think it is fine if replays stop working after some time; include a version number in the replay file and check it before attempting to play it back. Games like Warcraft have always had replays and obviously when there was a big patch, replays from older patches stopped working. That is acceptable.

    Also, couldn't you in theory have a speed setting for the replay, as well as forward and back buttons to go between actions? And doesn't the game already have access to everything you need in order to recreate a game state? Literally everything is in the battle log except for character position... How hard would it be to, for starters, just have a way to export a text file with every action taken and card drawn? I don't know, but I feel like replay functionality is eminently doable! :) and if you guys don't want to build it into the game, as long as you can export the battle log AND position data, I could definitely write a web app to play through the game...
     
  14. This is almost identical to how I've seen it done before with one other game. We even used Challonge :), which is an awesome service and perfect for tournaments like this one. I mean it would be crazy not to use Challonge.

    If I remember correctly, the first round was sudden death where you fought a random opponent and the loser was instantly kicked out. This was because it's a nightmare to manage a 32-player double-elimination tournament. By having that sudden death round, you can do a 16-player double elimination which is easier to manage, and most importantly, faster.
    After the sudden death round we had best-of-3 matches, but I don't think it's a good idea in Card Hunter because the matches last so long. I would keep all rounds best-of-1.

    And the answer the question of whether or not a system like this works, all I can say is that we ran it for over a year without any major issues. Many people here seem to be keen to have referees but I don't think they are crucial. People rarely cheat if you design the rules so that cheating is not easy (for example minor penalties for disconnections even if you have a "real" reason).

    I like single- and double-elimination tournament more than the swiss style tournament we had because instead of having random matches here and there, we have this one big epic event where winners are crowned on the spot.
     
  15. Flaxative

    Flaxative Party Leader

    While I agree with Farbs that running such an event as a contrast exercise would be interesting, and would be more than down to do so, it is not the kind of event that would appeal to me in the long term. I just don't see enough differentiation between it and normal ranked play. You can already play best-of-1, no deck restriction, effectively 'sudden death' matches anytime with anyone on the servers. The challonge bracket would just be a nice little hat to put on top of an environment that already exists. Maybe people like being called champions for playing games like this, but there's already the Elo system which even visually crowns the top players. And, of course, winning in ranked gets you rewards.

    So I'm a bit skeptical of how much such an event really brings to the table, but like I said, I am interested to try it out.
     
    Bearson Onyx likes this.
  16. We ran those kinds of tournaments for over a year and they were very popular weekly gaming and social events. The way I see it, there are two appeals:
    1. Each tournament would have a theme. In practice this would mean some kind of deckbuilding restriction, like "Priests Only", "Naked Characters" (no items), "Drawback Items Only", or even "Peasant" etc. etc.. This by itself would keep things fresh and interesting because there are a million different themes we could come up with.
    2. We would need some kind of special reward from BM. This could be pizza (a small amount to 1st and 2nd place winners), or maybe an item or something. Maybe an epic item of choice for winner, and rare item for 2nd place?
    Unless both of those two things happen, I don't think this kind of tournament would have much appeal, because like you said, it would be too close to regular PvP.
     
    Bearson Onyx likes this.
  17. Megadestructo

    Megadestructo Shark Card

    Hmm...all very good ideas...
     
  18. Kalin

    Kalin Begat G'zok

    I've been thinking of suggestion this one, but I'm seeing some problems: Elves would have a huge advantage in grabbing the VS early, and if neither player uses elves the games would take forever.
     
  19. Flaxative

    Flaxative Party Leader

    Dwarves could take SO MANY MORE weak strikes before dying!
     
  20. ElShafto

    ElShafto Goblin Champion

    NAKED DWARF RUMBLE



    ... wait, this isn't that -other- forum I read, is it?
     

Share This Page