Thursday 11 January 2024

Homemade Cube Part 15: Update


This is basically now just a blog diary of the homemade cube progress and is much more of a thing for me than anyone else. As such, only those that are interested in following the design process of this project are going to get much out of reading more here. I have had my second dose of new cards for coming up a couple of months now. I will upload the full spoilers of this second pile of new cards soon. 

We are in the sweet spot time window for testing these new cards and the changed meta, where you have a good idea of the situation but there is still plenty to learn and root out. It is no longer the blind fumbling about stage, but it remains a long way from being any kind of solved as a format. This dose of extra cards was about 200 new card designs aimed at solving issues, and about 200 revisions to old cards, of which nearly half were just aesthetics or grammar. Of the 100 or so rebalanced cards almost all were nerfs. There was the odd buff and the odd rework but for the most part things needed toning down. Turns out I am very good at knowing what I need to add to a card to make it playable, but not so good at knowing where to stop to keep it fair. This is of no surprise given I spent over a decade searching for cards that reach a minimum of power level rather than those not exceeding a maximum. It is also kind of useful as far as problems go. In order to test a card one has to play it. A card too weak is hard to spot and takes a lot longer to do so as you are spotting it by an absence of action. Consequently having not seen much of the card you have no idea why it is bad or by how big of a margin. At least if a card is too good you get a good idea of why quickly, and can then cut it if it is being oppressive. 

I had managed to get the size of the cube down to about 600 before the new cards hit. Wanting to keep it below 720 cards I did some fairly aggressive culling. Mostly just boring stuff that was fine but well understood. I have strongly felt that a 540 cube was a near optimal size for quite some time now but it turns out this is somewhat specific to the context of a cube using the real pool of magic cards and the bottlenecks and power gradients found therein. If you make a power gradient less steep and widen bottlenecks the size of your cube can expand as much as you like. There are some perks to larger sizes but on the whole I think closer to enough is better than way more, mostly just from a logistical point of view. A smaller cube is an easier shuffle and sort etc. It is also more accessible to newer players. With this in mind I didn't add all the new cards at once and drip fed additions in along side cutting things to keep size down. I am running a little below 680 at present and would like to get that closer to 600 in time but I am not in much of a rush and no longer think it is relevant to aim at 540 as before. Practically I think I am going to try and settle on 612 as that is largest single order print size the company I am using prints to. That means it is going to be nice and convenient to get a whole working cube printed off in one go and will make for a cool gift to a couple of mates.

The new cards, while a long way off the quality of the first lot, have done the job at solving the issues I was having with the format to begin with. There is now enough top end power and games are ending more consistently within a reasonable time frame. The first go was good, we got closer than I was expecting to the mark, but it was still only about 80% of the way there. This new batch seems to have gotten us to over 90% of where I was hoping to get to. As such the third shot at new cards and reworks/rebalancing should take us to over 95% which I am more than happy calling a finished product. I'll be able to cull to the best 612 once I have had testing on the third release. Any changes from that testing can be implemented on those 612 best and then I can consider it a completed project. The main upshot of all this is that once completed I will stop going on about it quite so much! I will likely do a bit of tinkering and add a few new ideas and tweaks in over the years from that point but it will not be taking all my attentions and will not dominate what ever remains of the blog by that point!

Testing windows will be increasingly long. The first lot only took a couple of months to identify the issues, and then no more than a couple more to have done enough design work to feel like they might be solved, this time round it is about half the pace, and so I don't expect to be in any place to make the third order until summer this year, and from then it is likely going to be winter again before I am settled on a final 612 and that final load of cards are sufficiently tested. Assuming that prediction is on point it means it will have taken about 18 months to turn over about a thousand cards. Much as that feels impressive given the small scale of the operation, it does not feel repeatable. This first go is milking years of idle thoughts on the matter that subsequent efforts cannot lean on. 

There is also a good chance the Art Cube project concludes within this time frame and puts it on hold for a while. We are now at about 900 cards (I was aiming at 1000 before we do a pre-print cull, although I now want more like 1200), and I am at around 80% viewed of the artists I was wanting to look through. As such there is every chance this is something I print before the summer. I am not even going to predict how long it might remain in the limelight for. Disregarding the Art set however I think we are looking at three or so months of testing where we are now, a month in limbo while we wait for the third and final big order, then a good six months of testing that.

There are a bunch of cards I already changed once and have now deemed they need a second change. Typically these have been the more complicated cards like the Class cards which just have more moving parts to understand and balance. I reckon if I can't get a card right in 3 attempts it is probably best to move on. That being said, we are still finding typos from the initial print run, cards that have been played repeatedly. There was an Edict called Edit and a Yavimaya card spelled Yayimaya that we only just spotted last time we played. Both great cards seeing a lot of play and both there since the beginning. I also have some vey literate and pedantic friends who have been pretty eagle eyed at spotting other errors. All that to say I expect these sneaky subtle errors will outlast any balance issues there might be on cards! 

My design objectives have morphed a little now. I was aiming for clean, simple, balanced cards in open design space in the beginning. I wanted to represent every aspect of the game. Now I am far more focused on making every card a card that increases the fun of the format. It was almost over balanced, over fixed, to the point games were too long, a little too taxing on the brain, and a perhaps even a little dry like chess. Effectively I overshot my goal and discovered things I didn't know about Magic and game design. The new Holy Grail of design for me is maximum fun. The first printing got a really nice framework of core set style cards with which to build upon with more fun, interesting, and complicated cards. I am in fact now working on some double sided cards. Much as I find them a bit overly complex and a bit of a phaff to play with it turns out that some of the most beloved cube cards are flip cards. If ones goal is to make the most fun format then it would seem remiss to exclude flip cards. It is mildly concerning as flip cards are more complex, with more moving parts, and I have clearly demonstrated to myself that those are the types of cards I am worst at designing. Further to that I have had almost no experience of playing with battles and so I expect my designs for those to be even more off the mark.

Speaking of card types I have not played with the fortifications I designed have been interesting. They are like little land equipment. I had about three times as many designed as I actually printed off and tested but I was so unsure about them I kept it to the most appropriate of the bunch to start with, the ones I was more confident of being viable. Mostly those have been a success and with what has been learned from them I hope to bring a couple more into the fold.

Companions are a card type I am working really hard on. They are super fine lines as far as balancing goes but that isn't so much the issue. It is really all about making them fun, something people want to build around. I had both playable and unplayable companions but some of those playable ones did not excite and that was an issue. Companions in a cube setting allow you to tie together some narrower elements. They let you draft a synergy deck more viably as you will always have part of your tool kit available. As such I have tried to lean into this and allow otherwise overly synergy based narrow archetypes to exist as supported by a fun companion. Turns out there is a lot to say about companions and this is already an overly long article so perhaps they get a segment of their own. All sorts of quirky things unique to them crop up, like, a companion can be boring because of how you have to build to include it as well as the effect it has on the game. I managed some of both of those! I also managed to make some companions that support the correct cool narrow archetype in a colour pair but then failed to correctly offer the required support. I changed the green white companion into an enchantment based one which was great, but then I made it a value tool rather than a reach tool and consequently no one ever played it because it was either no use or total overkill. 

As far as the meta goes, I am quite happy with it. I would call it a format of Anthems and Wraths as they seem to be the two effects that dominate when trying to pin point the cause of a win. There are still areas to improve upon but I would no longer say there are problems that need to be fixed, which is probably a big part of why the testing and revision phase is slowing up, there is less urgency. 

Control is a little stronger than I would like but it isn't dominating. The format could use a few more things that are awkward to control, and that can then lead to a victory. I have plenty of the former but it isn't getting the job done quite so well as it is rarely also the latter. Planeswalkers are no longer busted but they are still generally among the top end of the power range. Decks that are slower and more controlling are best placed to leverage them, as well as the Wraths, which in turn trump the anthem based builds, and that is all feeding into this mild imbalance. I also gave blue a bit too much meat for the board, a few too many efficient dorks with a lot of stats. This let blue really leverage its control cards and throw its weight around. The colour really needs the limitation of being a bit thin and outclassed on the board. Blue was the weakest colour prior to the second printing, now it is the second strongest. Luckily it is not out by much at all, and by less than my normal cubes are typically too! 

The card draw side of things is very much where I want it to be. It is hard to run out of stuff to spend your mana on in most decks which is precisely what I was hoping for. Moving beyond the specifics of Wraths and Anthems there is a really obvious correlation between winning and mana development. There is high demand on the mana rocks and ramp, with green likely being the best performing colour, simply because access to more mana over time has strongly tended towards winning. Most decks are making most of their land drops too, all the way past turn six. The sub 16 land format I predicted has not come to pass but it is not too far off. If the threats were as they are in the normal cube but had the support and value mine offers then it might well be a 14/15 land format. The thing is that with games going longer and good value on offer, even the cheaper and more aggressive decks wish to carry on deploying land for quite some time. It feeds off itself as well. The more value you have, clues to crack etc, the more lands you want to deploy to do so, and in doing so, you draw more cards and with them more lands thus fuelling those extra drops. A 16 land deck in my cube typically feels like a 19 land deck in my normal cube in terms of being able to drop a land consistently each turn well past the midgame. 

Much as the new injection of cards solved the teething problems I was finding with the format it is very noticeably lower quality in terms of design. I spend ages preening through the cards in the first release and tried to make it as near perfect as I could. By the time the second lot were out I appreciated how easy it will be to amend mistakes, and how often I would likely be making more cards and really lowered the bar on my output. I was happy just churning out cards I knew were off on power level or a bit sloppy on design because I wanted to test them, see quite how out they are and the effects that had. You can learn more from a mistake after all!

A lot of the top end I made to ease that bottleneck was a little pushed, as was a lot of the stuff aimed at boosting mono coloured strategies. The latter has not proved a problem with the cards being fun and well received and bringing the mono decks to roughly where we want them. They were rather underperforming before hand. The mana sinks all over performed as well but this was down to my failing to appreciate game length difference. If things speed up a little such cards will cease to be issues and so I don't feel presently like I need to directly address them.

A bunch of the new cards are just not playing all that well. The catch-up mechanic doesn't feel nice and is a bit confusing. Some of the more pushed cards have been nearing oppressive on occasion on the old power level side of things. The heavy handed "fixes" didn't have the desired affect at all such as the sparkhunter cards as a knee-jerk reaction to the initial potency of the walkers. Going forward I will take a much more gentle and gradual approach to solving issues and I will tackle these problems at both ends at one rather than just hitting the one end real hard and hoping it works out! Overall I expect only about 20-25% of the new designs to remain in the finished cube which is a really low take up. This is especially impressive given how clearly the power level had raised between the two sets of cards. Despite having complained about power creep for so long I immediately went and did it myself and worse! Luckily I am still in the testing phase and can rein in the things that went too far. 

The main takeaway here is that things are going well, the project is really fun, much more illuminating that I had anticipated, being enjoyed by my whole group and not just me, and likely looking to be the main way we cube going forwards. Despite the relatively sloppy, careless and rushed second batch of cards and their low take-up rate they have done the required job and we are very happy playing a format with no issues and that we all have rather more agency than before to improve as we see fit.

No comments:

Post a Comment