Page 62 of 88 FirstFirst ... 125258596061626364656672 ... LastLast
Results 1,221 to 1,240 of 1741

Thread: [Primer/Deck] Burn

  1. #1221
    Member
    Lyle Hopkins's Avatar
    Join Date

    Feb 2013
    Location

    Canada
    Posts

    117

    Re: [Primer/Deck] Burn

    Because of Miracles' prevalence in the meta-game right now, I'm curious what everybody is doing to combat this match-up. I was watching Joe Lossett's stream the other day and he mentioned that Legacy Burn should be playing four copies of Exquisite Firecraft.

  2. #1222
    Member

    Join Date

    Oct 2013
    Location

    The Naki, NZ
    Posts

    123

    Re: [Primer/Deck] Burn

    Quote Originally Posted by Lyle Hopkins View Post
    Because of Miracles' prevalence in the meta-game right now, I'm curious what everybody is doing to combat this match-up. I was watching Joe Lossett's stream the other day and he mentioned that Legacy Burn should be playing four copies of Exquisite Firecraft.
    In the maindeck, I have 1 Sensei's Divining Top and 1 Barbarian Ring. Barbarian Ring is pretty good as the last source of damage underneath a countertop lock, and Sensei's Divining Top is very efficient at digging for more Burn spells.

    In the sideboard, I have 3 Exquisite Firecraft, 2 Vexing Shusher, and 2 Pyrostatic Pillar which can come into the match-up. Exquisite Firecraft is self-explanatory, and probably the best card in the match-up. Vexing Shusher is nice on Turn 4-5, assuming that they're not on the Legend build, where you can just burst through if necessary. Pyrostatic Pillar is useful to try and get under Counterbalance with, especially considering how often Miracles players like to cantrip or tap and replay Top.

  3. #1223
    Global Moderator
    mistercakes's Avatar
    Join Date

    Nov 2009
    Location

    Copenhagen
    Posts

    2,274

    Re: [Primer/Deck] Burn

    has there been any sampling done with the new 4RR bedlam reveler? i know a lot of the issues with burn is that they "flood out" or maybe they had to use some resources on creatures/planeswalkers.

    has anyone tested him at all? i would like to, but don't have burn built as i'm missing the guides. (got rid of mine a while ago)

    (not to hear arguments about why he's bad, i just want to know who has actually tested this guy...thanks)
    -rob

  4. #1224

    Re: [Primer/Deck] Burn

    Quote Originally Posted by Lyle Hopkins View Post
    Because of Miracles' prevalence in the meta-game right now, I'm curious what everybody is doing to combat this match-up. I was watching Joe Lossett's stream the other day and he mentioned that Legacy Burn should be playing four copies of Exquisite Firecraft.
    He's 100% correct. I started off SB'ing them, I run all 4 MB now. It's a very strong card.

  5. #1225
    The Original

    Join Date

    Jul 2009
    Location

    Austria
    Posts

    34

    Re: [Primer/Deck] Burn

    The last tournament i played in i had 2 Exquisite Firecraft in the main and 2 in the sideboard. The thinking behind that is that they replaced the Sulfuric Vortexes because lifegain isn't that often seen in Legacy but counters are. And if you cast Vortex on turn 3/4 it takes 2 turns to deal the same amount of dmg as Firecraft does and at this stage the game should be over anyway. So i deals the same amount of dmg just slower and counterable/destroyable (sideboardhate against enchantments - Eidolon).
    From the tournament i can say it was totally worth it and i was thinking about putting all 4 mainboard.
    Drakes claim to be dragons-until the dragons show up.

    The dust can count only to 50.

    Angels attacked them,............ with swords.

  6. #1226

    Re: [Primer/Deck] Burn

    Some more updates to my program. Going to keep this one as brief as I can. I finally got around to adding Searing Blaze but due to some issues that I found extremely interesting I couldn't just take the quick and dirty approach to it. I actually had to do things the right way and create a targeting system. This comes with a few benefits. Most notably, Lightning Bolt is actually Lightning Bolt now rather than mimicing Lava Spike. I was even able to do it in such a way that the logic is generalized, which means one function of targeting logic can equally let Burn, Delver, and Miracles all find targets for their spells. The downside is it made card creation a little more unwieldy. Right now I can make cards that target, creatures, players, or creatures and players or not target at all but i cannot make cards that target two creatures. This means no Electrolyze. In theory I can do Searing Blood but I don't really have an event trigger system so it would require custom logic (which I don't want to do) rather than just parsing targets and numbers on the card. Searing Blaze however was doable. Based on the results, I'm glad I worked Blaze in but I don't think I'm going to need to work Blood in.

    The list I have right now is not the optimized list against the opposing AI, because I think optimizing for a bad deck is a poor idea. Instead I'm just looking at various card values. With my current list (posted below) over 10,000 games Searing Blaze looked optimal at 4 cards (score of 0, so perfectly balanced).

    Firecraft is looking pretty damn good as a mainboard 4 of as well. It does work out poorly in turn 3 games where it's over represented as a 4 of, but in the turn 4 games it's only slightly high and in turn 5 games it's under represented. Depending on the expected meta I could see an argument here to only play 3 MB if you expect things to be fast.

    Guide and Swiftspear are starting to climb the charts of over representation now that there's an opponent with blockers and the ability to kill them. According to the numbers Guide is a little better than Swiftspear, but Prowess gets exponentially more powerful (just think about it, it's way better when you trigger on two creatures at once) as you put more of it in your deck so the real weaker card might be Guide.

    Lands show up as under represented still.

    Finally, here's the list and it's worth noting that this is the first list to go through my program (ever) that had more turn 4 wins than turn 5 wins, despite the fact that the opponent has blockers. Searing Blaze is just that good.
    22 Mountain
    2 Barbarian Ring

    4 Eidolon of the Great Revel
    4 Monastery Swiftspear
    4 Goblin Guide

    4 Price of Progress
    2 Fireblast
    4 Lightning Bolt
    1 Rift Bolt (program actually just uses 5 Lightning Bolts right now)
    4 Exquisite Firecraft
    1 Sulfuric Vortex
    4 Searing Blaze
    4 Chain Lightning

    One last comment, over the past few days I've been thinking about how to implement Bedlam Reveler because he's new, topical, and I bet a lot of people would like some extra data before Louisville in 5 months. Bad news. I'm not sure I can implement him. I'm going to try, but given my limited knowledge of AI programming I'm not sure how to write a variable cost. It's a really difficult concept to explain to a computer. I have some ideas, one is a really ugly way to implement it but would work. The other would probably require expanding my trigger system but would be a little more sustainable. If I implement it that's the direction I'm going to take.

    If I'm being honest with myself though, the card I really want to know about is Collective Defiance. I like this card more in Modern but I think it has some SB potential here against Miracles and combo. Sadly, modal spells are way beyond me. Modal spells are 500x harder than variable mana costs. Spells with both of those combined are probably impossible given my current approach.


    Quote Originally Posted by Gollus View Post
    The last tournament i played in i had 2 Exquisite Firecraft in the main and 2 in the sideboard. The thinking behind that is that they replaced the Sulfuric Vortexes because lifegain isn't that often seen in Legacy but counters are. And if you cast Vortex on turn 3/4 it takes 2 turns to deal the same amount of dmg as Firecraft does and at this stage the game should be over anyway. So i deals the same amount of dmg just slower and counterable/destroyable (sideboardhate against enchantments - Eidolon).
    From the tournament i can say it was totally worth it and i was thinking about putting all 4 mainboard.
    I agree with that logic. They're good on their own though even without the lifegain. They're fast, and while not the most mana efficient they are very card efficient. I think that's something we take for granted too often when playing Burn. One of the approaches to the deck is that we're a combo deck, we try to resolve 7 3 damage spells to win the game. If we can add just a single 4 pointer to that list we're trying to resolve though, it's an entire card less that we have to throw at our opponents face. Would you rather try and resolve 7 spells for 10 mana or 6 spells for 11?

  7. #1227

    Re: [Primer/Deck] Burn

    A few more results, I wanted to get a large amount of data in with the AI so I ran a 100,000 game set while I slept last night.

    It had an average game length of 4.678 turns.

    Turns games ended on
    T2 - 3 (I am confused by these)
    T3 - 4250
    T4 - 41458
    T5 - 36238
    T6 - 12884
    T7+ 5167

    One of the things I really wanted to test here was to evaluate Swiftspear. Previous goldfish games were coming in with Swiftspear ranking slightly better than Goblin Guide even though the damage queries show it to be strictly worse in damage output. I have no explanation for this yet. Swiftspear ranks as the better card, despite putting out less damage and being able to swing less often and the Guide downside can't be the explanation because my Guides don't have the attack trigger on them (yet).

    Anyways, the observations from this. I added a T6 comparison to the chart to weight for an optimal deck. Similar results as before. Lands are too low, 1 drop creatures are too high. I might try out a Marauder.

    My next project will be cleaning up the code and adding proper documentation, it's been a while since I've posted it publicly so once it reads as something fit for human consumption I'll be doing that for anyone interested.

    Worth pointing out though is that Searing Blaze came in with 4 as the perfect number in all games from T3 to T6 with an opposing deck that has 12 creatures.

    If anyone has requests for decks I can try running them.

  8. #1228

    Re: [Primer/Deck] Burn

    You're posting update after update after update on huge sample sizes but as mentioned the high land count is extremely suspicious. You're running control-deck level of lands, which is absurd given this is burn we're talking about.

    FYI, and since you mentioned his experiments (which yours are similar to), FK did converge to the conventional wisdom of 20 lands for an optimal aggro deck. Now, that's for an optimal goldfishing (no opposition), but still.

    This kind of approach via numerical simulation is most certainly justified and valid enough for "simple" decks like Burn/RDW. Yet as with any tool, results can only be useful if the tool behaves as intended, and it's obvious something is not right when you can announce up to 26 lands.

    Edit : the T2 wins you are confused by are another indication that some things are not right. This particular point might be of a totally separate issue to the land count one, but it serves as a pointer that there are some significant bugs still in the tool.
    Quote Originally Posted by cdr View Post
    140x Relentless Rats
    Quote Originally Posted by Ben Bleiweiss
    I wish that Wizards would have just gone ahead and done away with the Reserved List entirely. It is nothing but a blight on the game and one that long outlived its purpose. [...] I am wholeheartedly in favor of getting rid of the Reserved List and reprinting higher-dollar staple cards from EDH and Legacy. Pete Hoefling the owner of StarCityGames.com agrees with my point of view as well.
    - Ben Bleiweiss, SCG General Manager, Feb 2010

  9. #1229

    Re: [Primer/Deck] Burn

    Quote Originally Posted by ParkerLewis View Post
    You're posting update after update after update on huge sample sizes but as mentioned the high land count is extremely suspicious. You're running control-deck level of lands, which is absurd given this is burn we're talking about.

    FYI, and since you mentioned his experiments (which yours are similar to), FK did converge to the conventional wisdom of 20 lands for an optimal aggro deck. Now, that's for an optimal goldfishing (no opposition), but still.
    His other posts came to different conclusions on land. Most notably, once he got away from a linear system of 1 mana 2/x, 2 mana 4/x, 3 mana 6/x. In that situation a mix of 1 and 2 drops actually performs optimally. You don't gain enough card equity from having more power in one card because the game is ending before you empty your hand. That's not quite the case with Burn though because there's more one shot effects and fewer recurring ones.

    If you drop the curve the land count goes down, an all Bolt deck wanted I think it was 21 lands. but after having been playing with this stuff in Burn for 2 months now, and having played with it in other decks I've attempted for about a year now... the results have always been the same. Faster decks want more land, slower decks want fewer land with some card draw. That sounds backwards to how most people build decks but it actually makes perfect sense. With more lands you have a better chance of curving out, and when you curve out you have more options, gain more tempo, and can play your cards faster. It also leads to having more cards in hand overall because you mulligan less. Yet with fewer lands, your draws (particularly aided by card draw) smooth out and still provide a steady 1 land/turn.

    Another way to look at this, is that his deck had the luxury of playing all 1 and 2 drops. Burn these days though is often times interested in 3 drops, if not in your MB then in your SB. Cards like Ensnaring Bridge, Exquisite Firecraft, and Sulfuric Vortex are very good, and there's little to no debate on that. Once you accept that you will be running some 3 drop, you then have to add enough lands to reasonably cast those 3 drops and 20 just doesn't do that. If you force the simulated deck to include some 3 drops, you'll see your land count rise by quite a bit.

    One thing I do need to do though, is do the same thing I did early without an opponent and churn through decklists using every land count from 18 to 30 and then looking at what land count is fastest and not necessarily the most well balanced (which appears to be at 27 lands). In fact, I'll see what I can do about doing that tonight. But in my rewrite for an opponent I lost the ability to easily que up a bunch of decklists at once so I'll have to tinker a bit for that.

    This kind of approach via numerical simulation is most certainly justified and valid enough for "simple" decks like Burn/RDW. Yet as with any tool, results can only be useful if the tool behaves as intended, and it's obvious something is not right when you can announce up to 26 lands.
    I'm not sure I agree, and this doesn't just come from this program. It comes from experience building decks, I've always been the sort of person who plays 1 or 2 lands extra (which raises the question of if I'm right, or if it's a bias inherent to my approach). Having the mana to play your cards is the single most important thing in the game. I'm not actually playing 26 lands right now, but I have played 26 lands in aggro decks in the past and it worked out well and probably will again in the future. Usually what you're looking for in those last few lands are mana sources that are also mana sinks. Unfortunately the mana requirements in Burn (lots of RR) prevent that from easily happening. 26 Mountains is probably less than optimal but 26 lands could be, in fact as I write this I'm testing a 27 land.

    Also, I think this approach can work for any linear deck eventually. Burn is one of the easiest starting points but decks like Jund, Maverick, and Delver aren't much different. It's when you get into non linear decks (anything tutor heavy) and combo decks that it ceases to function, though I suspect I could make AI's specific to each combo deck with enough time and practice.

    Edit : the T2 wins you are confused by are another indication that some things are not right. This particular point might be of a totally separate issue to the land count one, but it serves as a pointer that there are some significant bugs still in the tool.
    It's not an issue that things aren't right, it's just an issue with looking at the results (seeing a T2 win) and not being able to see any lines that lead to that result to explain it. I ended up looking a bit deeper into those games, and discovered the lines. All three were slightly different but basically involved something like this sequence.

    Burn T1 Goblin Guide
    Delver T1 Fetch, Gitaxian Probe, Gitaxian Probe
    Burn T2 Flame Rift
    Delver T2 Fetch, Flame Rift
    EoT Fireblast

    That sequence, or something very similar came up 3 times in 100,000 games which is where the T2 wins came from. You could chalk that up to either a poor decision in deck construction giving it a PoP/Flame Rift type spell, or a poor decision on the AI's part to cast it in that situation (I don't know what other cards it had to pick from when deciding on a play). It's an example of where interaction actually speeds things up. With Delver which burns it's life total in various ways between Probes, Fetches, and Flame Rifts most games are actually faster than a straight goldfish.

    -------------------------------------------------
    Edit: I looked deeper into land counts and win turn using 1000 games per land count. I'm going to rerun the test tonight with 10,000 games. The 1000 game dataset suggested little to no effect on the average win turn between 17 and 34 lands.

    Here's the chart of win turns
    http://imgur.com/5PHPI78

    In the dataset, the fastest deck was 4.51 turns at 22 land and the maximum was 4.646 turns at 23 lands. Everything between 17 and 34 lands fell within that range. Therefore the conclusions are either that the sample size is too small to say anything (and if you look at that chart, there's no curve to suggest a range for the optimal count), or that it simply doesn't matter. The sample size can hopefully be ruled out soon.

    ----------------------------------------------------
    More edits: The reason the 17 deck set showed very little difference with land counts is because of a bug I had. Nothing to do with the logic of how it plays games, but when I would reset my deck stats at the start of the turn it would reset the deck to the initial configuration. So even though I fed it 17 decklists to try each one 1000 times it tested the same decklist 17,000 times.

    That's fixed and the 10,000 game set is running tonight. Before setting it up I ran a 1000 game set. I don't see any point in posting those results since a better dataset will be available tomorrow. Suffice it to say, there's an actual curve on the win turn again, though it's somewhat minor (about .4 turns between the fastest and slowest deck). Because if that small difference between the fastest and slowest, I went through the extra trouble of tuning each list to it's mana curve. This results in 17 different lists rather than 17 very similar lists aside from mana count. I'm hoping it works out well since it took something like 3 hours to set up as it's a rather labor/time intensive process.
    Last edited by Brael; 08-11-2016 at 01:29 AM.

  10. #1230

    Re: [Primer/Deck] Burn

    Ran the program, looked at the results. Here's a chart of the average win turn, only taking into account games won/tied.

    http://imgur.com/Hoec9ku

    The curve isn't quite as clear as I was hoping for, but oh well, it's still clearly there, and you can see where the av

    The fastest deck was the 19 lander, and the next fastest was at 23 lands, and everything else was atleast .1 turns slower than that. Oddly, 26 lands proved to be the slowest by a big margin (I was expecting it would be 33 lands). The 26 land deck also had the lowest win rate of all decks, it was so far out of line that I reran that deck two more times, each time tuning the deck to something better and each time returned the same result, it was losing about twice as much as the other decks. I don't really have an explanation to that, but it's odd and consistent.

    So I ran the tests again and this was the chart
    http://imgur.com/qnlr71I

    Basically the same results with 23 performing best. 19 was good but still not quite as good. The over 30 lists repeated their odd decline in turns. My best explanation for that is in those higher lists I started playing a bunch of Keldon Marauders and they still seem to be quite good even though we've kind of forgotten about them.

    The difference between 19 and 23 lands is basically Firecraft. Here's the lists those two decks used (for now Searing Blaze is always 3 damage which has a slight effect):

    19 Mountain

    4 Eidolon of the Great Revel
    4 Monastery Swiftspear
    4 Goblin Guide

    4 Lightning Bolt
    4 Chain Lightning
    4 Rift Bolt
    3 Lava Spike

    4 Price of Progress
    2 Rift Bolt
    4 Searing Blaze
    1 Exquisite Firecraft
    2 Fireblast

    1 Pyrostatic Pillar


    22 Mountain
    1 Barbarian Ring

    3 Monastery Swiftspear
    3 Goblin Guide
    4 Eidolon of the Great Revel

    4 Lightning Bolt
    2 Rift Bolt
    4 Chain Lightning
    4 Price of Progress
    2 Flame Rift
    4 Searing Blaze
    4 Exquisite Firecraft
    2 Fireblast

    1 Pyrostatic Pillar

    With the 23 land decks performing best in both average turns and number of games won against Delver as an opponent , that's probably what I'll focus on trying to optimize for a bit.

  11. #1231

    Re: [Primer/Deck] Burn

    Quote Originally Posted by Brael View Post
    A few more results, I wanted to get a large amount of data in with the AI so I ran a 100,000 game set while I slept last night.

    It had an average game length of 4.678 turns.
    ....
    I'm a little unclear. What is the simulation here - Burn AI vs Delver AI, or Burn AI goldfishing or something else?

  12. #1232

    Re: [Primer/Deck] Burn

    Quote Originally Posted by rufus View Post
    I'm a little unclear. What is the simulation here - Burn AI vs Delver AI, or Burn AI goldfishing or something else?
    Burn vs Delver AI's

  13. #1233

    Re: [Primer/Deck] Burn

    Added a new card to the card pool today, Forgotten Cave. On first glance this isn't really something Burn is interested in because we don't want to use CIPT lands, but I'm always on the look out for cards that double as mana sources and mana sinks, so I was curious.

    The 27 land test deck with 4 Caves substituted in for Mountains performed about equally to the 22 land test deck, with none. 25 land with 4 was very close to 23 land with 0, but slightly worse. So as people probably could have guessed, it just wasn't good enough. It was much closer than expected though so I'm probably going to look into this a bit more and see if there's an optimization I can make to the list somewhere.

    In mono colored decks fetchlands have been poor performers, so having another land based way to get some cards into the GY for Lavamancer/Barbarian Ring could be nice. It will require a build designed to take advantage of it rather than a generic list though. I don't expect much, because as I said before CIPT is really not where we want to be, but I'll likely play around with it over the next day or two and see if anything interesting shows up. Conditional mana sources could be part of what we need to make Bedlam Reveler work.

    Sadly, my card rating spreadsheet that has so far done a great job of identifying correct numbers for each card, doesn't quite work with this one (in the games we want to optimize it's almost always being played as a land, which basically just makes it a strictly worse Mountain). That system works by comparing how common a card is in our opening hand to the average we should see on a given turn. If it shows up above average, then it's more powerful. That doesn't actually work with something that's intended to be played out later in the game, so I'll have to retool the system a bit to see if it says anything.

    It's also very possible I'm just wasting my time on a card that's already been rightfully dismissed, but I'm curious. If anyone can suggest any other interesting cards that might be out there, I'll try and take a stab at quantifying them assuming it falls within the realm of what my software can either do already, or do with some small tweaks.

  14. #1234
    Global Moderator
    mistercakes's Avatar
    Join Date

    Nov 2009
    Location

    Copenhagen
    Posts

    2,274

    Re: [Primer/Deck] Burn

    I've been working on a bedlam reveler burn list for a few days now. going to test it with some fetch lands today to see how it changes the deck.

    will post list with my findings later today. (in eu right now)
    -rob

  15. #1235
    The green Ancestral
    ESG's Avatar
    Join Date

    Mar 2010
    Location

    Seattle, WA
    Posts

    1,308

    Re: [Primer/Deck] Burn

    Quote Originally Posted by Brael View Post
    In mono colored decks fetchlands have been poor performers, so having another land based way to get some cards into the GY for Lavamancer/Barbarian Ring could be nice.
    Or just don't play those cards if your deck has no fetchlands.

    Quote Originally Posted by Brael View Post
    It's also very possible I'm just wasting my time
    Going to stop the quote there. You have a lot of enthusiasm for programming, but I fail to see what meaningful results can be drawn from your experiments. You aren't testing against real opponents, and your simulations are, by your own admission, lacking. I want to give you the benefit of the doubt, but when you've gotten to Forgotten Cave, the posts are starting to seem like spam. Well-intentioned spam but spam nonetheless. What would be more helpful is if you posted your findings from playing the deck at real tournaments.

  16. #1236

    Re: [Primer/Deck] Burn

    Quote Originally Posted by ESG View Post
    Going to stop the quote there. You have a lot of enthusiasm for programming, but I fail to see what meaningful results can be drawn from your experiments. You aren't testing against real opponents, and your simulations are, by your own admission, lacking. I want to give you the benefit of the doubt, but when you've gotten to Forgotten Cave, the posts are starting to seem like spam. Well-intentioned spam but spam nonetheless. What would be more helpful is if you posted your findings from playing the deck at real tournaments.
    There's just not that many locally. I get a 3 round Legacy tournament roughly every other week and I split those between multiple Legacy decks (though when school starts back up, we might be able to get our weekly tournaments going again as people come back into town). I get games in where I can, but that can only tell you so much, and I've posted about those and had good results. A string of 15 or 20 games where I win 2/3 of my constructed matches just doesn't mean anything though because that's the results I get with any of my decks in Legacy, and the sample size just doesn't exist to glean any real information unless something just clearly doesn't work. I live in southern Ohio so it's reasonable to travel to the majority of big tournaments in the US (especially SCG's) but my schedule/budget just doesn't allow it, I prefer to use that money to buy cards instead. Though, there's a good chance I'll be going to Louisville (but I haven't locked on a deck yet). On top of all that, anecdotal evidence just isn't reliable.

    Forgotten Cave is there just to try. My tone may not have been clear, but I'm not expecting anything from it results wise. I think it would have a potential spot if we had lots of 3 and 4 mana spells to cast, but we don't. There's very few 4's that seem worth it. When I looked through the list as part of trying to optimize the higher land count decks Collective Defiance and Fiery Confluence were the only ones that seemed remotely interesting, and I'm pretty sure they're not good enough.

    Anyways, the card is probably more interesting to me than it is to you because while I'm testing Burn I'm also making a note of how it plays into general manabase construction. Maybe it's place isn't here, I'm going to play with it further to decide that though. I'm just going through the various possibilities. The strongest T4 hands have been the ones that hit 4 lands, but obviously if you hit too many lands your hand gets weaker. So I'm trying various things that can let you hit your land drops but still have the needed number of cards to kill the opponent. Fetchlands don't have much of an effect here (they have some, but at a high life cost), and Barbarian Rings are of some but limited value, so I'm looking at other possibilities. They may or may not pan out after testing. It's really no different than having tested Cantrips, Magma Jet, and SDT in the past... and none of those worked out (though I could have sworn from my paper results that SDT was working.... that's the difference between a tournament report and recording a million games to evaluate a card).

    I think you misunderstand the point of having an opponent as an AI. It's not to create the highest quality of play possible, even as a goldfish it doesn't do that. The point is to have something that's capable of blocking, interacting, and counter attacking. Recording 1000 games of Jon Finkel vs LSV will give you a high quality dataset, but high quality plays aren't what's important. Seeing how the cards react in the majority of situations is, and for that something less skilled works fine as long as you play enough games, because my goal is ultimately to create a deck skeleton and evaluate some cards against each other, not a specific list. My biggest limitation on this front though is simply having faster hardware. I say the AI isn't very skilled, but when I say that I'm judging it by the range of possible player skills, I would put the AI up against the DotP AI any day.

    As far as spam goes, that may be true, it's even something I take into consideration when posting, but part of discussion is bringing up cards that were never before considered. Regardless though I can stop posting about that card if you wish. My classes start back up in a week anyways and I have to comment everything and make it presentable to make the code available to all again, so once that happens this project won't get any work until winter break because I just won't have the time.
    Last edited by Brael; 08-14-2016 at 01:09 PM.

  17. #1237
    Global Moderator
    mistercakes's Avatar
    Join Date

    Nov 2009
    Location

    Copenhagen
    Posts

    2,274

    Re: [Primer/Deck] Burn

    been testing this for a few days now. here's the list i've been running on modo with moderate success.


    4 gitaxian probe
    4 manamorphose
    4 faithless looting
    4 lightning bolt
    4 chain lightning
    3 rift bolt
    2 lava spike
    3 fireblast
    3 exquisite firecraft
    3 price of progress
    1 searing blaze

    4 eidolon of the great revel
    4 bedlam reveler

    9 mountain
    2 scalding tarn
    2 wooded foothills
    2 bloodstained mire
    2 arid mesa


    here's the cards i've been testing in sb, nothing too crazy here:


    3 grafdigger's cage
    1 surgical extraction
    1 exquisite firecraft
    1 searing blaze
    1 price of progress
    1 ensnaring bridge
    2 red elemental blast
    3 smash to smithereens
    2 pyrostatic pillar




    this deck sacrifices a little bit of speed, but has been a blast to play. give it a shot if you like playing with the overpowered bedlam reveler. best play for this deck feels like casting reveler and with ability on the stack casting fireblast.

    the deck can reasonably cast reveler on turn 3-4, especially with a faithless looting. looting also functions as a way to fix the draws late game where burn usually falls apart.

    one thing to note is with this deck i don't mind firing all of my spells at their creatures as i know i will be able to refill later on. don't be afraid to use your hand at their guys....it's going to play different than bolting their face 7 times.
    Last edited by mistercakes; 08-14-2016 at 05:06 PM.
    -rob

  18. #1238
    Member

    Join Date

    Oct 2013
    Location

    The Naki, NZ
    Posts

    123

    Re: [Primer/Deck] Burn

    Quote Originally Posted by mistercakes View Post
    been testing this for a few days now. here's the list i've been running on modo with moderate success.


    4 gitaxian probe
    4 manamorphose
    4 faithless looting
    4 lightning bolt
    4 chain lightning
    3 rift bolt
    2 lava spike
    3 fireblast
    3 exquisite firecraft
    3 price of progress
    1 searing blaze

    4 eidolon of the great revel
    4 bedlam reveler

    9 mountain
    2 scalding tarn
    2 wooded foothills
    2 bloodstained mire
    2 arid mesa


    here's the cards i've been testing in sb, nothing too crazy here:


    3 grafdigger's cage
    1 surgical extraction
    1 exquisite firecraft
    1 searing blaze
    1 price of progress
    1 ensnaring bridge
    2 red elemental blast
    3 smash to smithereens
    2 pyrostatic pillar




    this deck sacrifices a little bit of speed, but has been a blast to play. give it a shot if you like playing with the overpowered bedlam reveler. best play for this deck feels like casting reveler and with ability on the stack casting fireblast.

    the deck can reasonably cast reveler on turn 3-4, especially with a faithless looting. looting also functions as a way to fix the draws late game where burn usually falls apart.

    one thing to note is with this deck i don't mind firing all of my spells at their creatures as i know i will be able to refill later on. don't be afraid to use your hand at their guys....it's going to play different than bolting their face 7 times.
    Mate, deck's awful.

    No Guide, no Grim Lavamancer, 4 too many copies of Bedlam Reveler/Gitaxian Probe/Manamorphose/Faithless Looting. Not a Burn deck.

  19. #1239
    Global Moderator
    mistercakes's Avatar
    Join Date

    Nov 2009
    Location

    Copenhagen
    Posts

    2,274

    Re: [Primer/Deck] Burn

    I might be able to find room for lavamancer, as I would be good in this list. I haven't tested him yet.

    I'm fairly certain based on your comment you haven't tried the list yet. that's fine. I'll post some videos of my matches and let them speak for themselves.

    edit: i'm sure there are some questionable plays, but this should show the power of this deck, and why i'm making concessions with probe+manamorphose+looting for reveler.

    first match vs miracles: 2-0 (this was in casual tourney practice room)

    https://recordings.join.me/v0VRvuuaw0SwmKoQMADI4Q

    2nd match vs junk nitfit: 2-1 (this was a 2 man event)

    https://recordings.join.me/jvosveXq5E6hqOFBl6pf1w
    Last edited by mistercakes; 08-15-2016 at 07:44 AM.
    -rob

  20. #1240

    Re: [Primer/Deck] Burn

    Quote Originally Posted by mistercakes View Post
    I might be able to find room for lavamancer, as I would be good in this list. I haven't tested him yet.

    I'm fairly certain based on your comment you haven't tried the list yet. that's fine. I'll post some videos of my matches and let them speak for themselves.

    edit: i'm sure there are some questionable plays, but this should show the power of this deck, and why i'm making concessions with probe+manamorphose+looting for reveler.

    first match vs miracles: 2-0 (this was in casual tourney practice room)

    https://recordings.join.me/v0VRvuuaw0SwmKoQMADI4Q

    2nd match vs junk nitfit: 2-1 (this was a 2 man event)

    https://recordings.join.me/jvosveXq5E6hqOFBl6pf1w
    I know I try some weird stuff, but your deck is too far out there for me. The goal of a deck like Burn is to kill on the fastest turn possible, and spending your turn doing something like Faithless Looting just isn't conducive to that.

    I remember a couple years ago a friend of mine had some success with a Delver/Storm hybrid that would play a typical delver game but use cards like Manamorphose and Probe to also chain into some large Grapeshots. It feels to me like you're trying to do the same thing, but then you don't have the support cards for it.

    On the whole cantrip idea, I think cantrips are in general pretty overrated. If you search back some posts here, I did some pretty extensive tests with Magma Jet, SDT, Probe, Serum Visions, Preordain, and now Forgotten Cave. They all turned out pretty much equal, which is that they were just a little worse than simply putting more good cards in your deck.

    4 Reveler definitely seems like too many. The card reads to me like a 2 of at the most because it's going to be prone to clogging your hand.

    I watched the 2nd match you posted, Nic Fit vs Burn, those happen to be my two most played Legacy decks (and the two I usually keep put together) so I know both decks well, and how they play together. I would caution against trying to glean too much info from that match. That Nic Fit player really screwed up, among other things Veteran's come out in games 2/3 and that player didn't do that. That alone strongly tilted the match in your favor, to say nothing of the many other errors I saw on their side of the table. Then again it was tournament practice which isn't really known for high level play.

    -------------------------------------------------------
    And a little more simulator info. Cave didn't test well enough, it's something I'll keep in the back of my mind in the future but it's great for T6ish decks, we're looking to end the game on 4 though so it's just not where we want to be, just like every other cantrip (worth noting though, it did outperform Gitaxian Probe).

    Instead, I went back to testing fetchlands. The previous fetch test in a goldfish scenario, and I wanted to see how much the life mattered. I tested several lists at 23, 24, and 25 lands (23 optimal, a couple extras because of thinning). Similar to what I found previously, 12 fetches suggest you want +1 land in the deck.

    The optimal count ended up being 24 fetches split 11 fetch/11 land/2 ring. The win rate went down slightly doing this but the average win turn also went down by a lot. The difference between 21 Mountain 2 Ring and the fetch manabase was nearly half a turn on average over 5000 games).

    The change in win rate was 82% with Mountains, 80% with fetches (the Burn deck is pretty favored in these games).

    So in the end fetches will increase your number of T3/T4 wins but slightly decrease your overall rate. Meaning, in a faster meta, play more fetches. In a slower one play fewer.

    ------------------------------
    One more bit of information on lands. Using the optimal manabase I listed above, I ran a 3000 game set and looked specifically at decks that won on T4. Here's the lands they had on the board, to give some indication of how they curved out (note, this information gets recorded after Fireblast is cast for the turn)
    0 Land - 3
    1 Land - 11
    2 Land - 151
    3 Land - 373
    4 Land - 652

    So out of 3000 games 1,190 ended with a victory on exactly T4, and of that about 61% of games hit 4 lands while 38% of games hit 3 lands.

    I think this is some good evidence in favor of my argument that Burn really wants to curve out and not just operate off of 2 lands.
    Last edited by Brael; 08-15-2016 at 06:55 PM.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)