Caverna resources

Caverna: Cave versus Cave (2017) – Accessibility Teardown

Game Details
NameCaverna: Cave vs Cave (2017)
ReviewMeeple Like Us
ComplexityMedium [2.55]
BGG Rank816 [7.11]
Player Count1-2
Designer(s)Uwe Rosenberg
Buy it!Amazon Link


27/4/2019 – As per discussion with Mike Crabb, the cognitive section has been reworked a little and the grade dropped.

Version Reviewed

English edition


Caverna: Cave versus Cave is a perfectly solid game but that’s not really, in itself, sufficient in this day and age. Competence is too low a bar to expect a game to clear, and while Cave versus Cave does so without meaningful error it still leaves me feeling notably unimpressed. I can’t dispute anyone that says it’s a good game, but I can’t imagine a circumstance under which I’d actively go out of my way to play it. For these reasons we gave it three stars in our review.

Ah, but you know the mantra by now – games on Meeple Like Us can be redeemed by a strong performance in a teardown. It’s not always the case that everyone is able to play the games I might want them to play. Even if they can, everyone’s tastes differ. Maybe this is your perfect game – maybe our review would have steered you away from something you could genuinely love forever. A game with good accessibility can end up being recommended again and again here on Meeple Like Us. Will Cave versus Cave thrive in that scenario? Let’s find out.

Colour Blindness

Colour blindness isn’t an issue for Cave versus Cave – the only place where colour is used as a differentiator of game information is in the blue (passive effect) and orange (active effect) rooms. These are colours that offer reasonable discrimination for all of our standard categories of colour blindness. For those with more specialised manifestations of colour blindness the effect of the tile is clearly documented on the front.

Colour blindness tiles

We strongly recommend Cave versus Cave in this category.

Visual Accessibility

The number of options with which a player will be presented will increase round by round – a new action square is flipped over every time a new round begins and the order of these is not wholly predictable. If someone plays Cave versus Cave a lot it’ll become possible to internalise what the different options do but they’re not particularly intuitive or clearly deducible by their text description.

For example, ‘undergrowth’ lets you take a room action from your board and gives you two wood. Undermining lets you take two room actions or excavate through walls. Furnishing gives you a point of food and lets you spend one food per a variable number of dwarfs to place a room. Housework does likewise but also gives you at a chance of a second placement. There’s a fair degree of role overlap and situational effect here. Memorising effects then will be possible if someone has a decent memory but it requires remembering the nuance of as many as twelve actions in the last round, together with those actions that are specific to a player’s board.

Player board

With that said though, the number of these options will diminish over the course of the round and it’s possible for a visually impaired player to ask if there is a path to the goal they want to achieve. ‘Are there any actions that let me excavate’, or, ‘What is the largest number of room actions I can take with the options available’. It’s not a good idea though because if your opponent knows you plan to build they’ll do everything they can to reduce the benefit of that to you. In any case this needs a player to have a firm idea for what they want to do as opposed to the more contextual decision making a sighted player would be permitted.

That’s important because Cave versus Cave is a game of optimising under difficult constraints – a worker placement game without workers and where even the basic mechanisms of play are available only unreliably. Close inspection of tiles is going to be necessary on a regular basis and they are often very information dense – they have names, victory point totals, cost to build, wall placement rules, and ongoing or activated actions they add to your player board.

Warehouse tile

Some of the action squares suffer from low contrast (particularly when dealing with costs) and visual silhouettes that are too similar for easy comfort. The ‘building a wall’ versus ‘razing a wall’ symbols are differentiated only by a plus and an arrow. Sometimes actions are separated by a slash which means ‘one or the other’ but visual impairments may make this non-obvious unless closely examined. The impact of making a mistake in your choices here is considerable.

Working out whether rooms are even feasible for a player’s board depends on an inspection of board against tile. The placement requirements can be rotated, and there is a physical representation of walls on the board. However, much of what goes into selecting a room depends on what you can do rather than what has been done. A player might for example build a wall so as to claim a tile that would otherwise be unavailable. This requires an understanding of game state that includes the number of actions available, the extent to which they are redundant and resistant to undermining, and a full understanding of all the options available to be purchased. Because rooms are defined by wall configurations with occasionally optional components it’s not as simple as deciding on wall layout and picking the room that fits. There’s a larger possibility space that needs to be taken into account and a wise player will be equally mindful of what an opponent might be plotting. There’s no point banking on a build action in your third turn if your opponent is going to use the only one available on their second. Timing is hugely important here and an appreciation of its impact is most easily obtained visually.

One excellent feature in Cave versus Cave, and I don’t think I’ve encountered this before, is that it gives you two redundant sets of markers for resources. One is large and tactile, offering physical differentiation by fingertip. The other is made up of small cardboard tokens that minimise space taken up on the board. This would have been a top-notch accessibility aide if it weren’t for the fact the flax and emmer tokens are a little too similar to be completely unambiguous in most circumstances. It’s still a welcome addition. Unfortunately, these are still placed with reference to a resource track that is indicated only visually and has no tactile differentiation between resources of one value and another. They crowd together because there are a lot of different currencies in the game and they are not given much room on the track. The end result is that if you do attempt to investigate game state by touch alone you’ll almost certainly end up changing it. It’s a lovely idea, not well implemented. Other games though should take inspiration from this.

Overall, we don’t recommend Cave versus Cave in this category.

Cognitive Accessibility

All of the information needed to play the game is presented with some kind of referenced token or visual reminder but some of it is peculiarly unintuitive. For example the number of turns you take is shown on the space ‘where the most recent action tile has been revealed’. It’s also shown by the number of dwarfs on the action board. That’s two pieces of redundant information except that the dwarfs are only lightly coupled to action squares and the number of the space you have just revealed will then be lost as the tile is flipped over. It’s not a major problem but it is kind of weird. To be fair, the action board also has walls that show where particular round counts begin and end but the manual makes a point of saying ‘the depicted walls have no impact’. It’s a little bit contradictory.

Memory too is stressed by the fact there are only a set number of rooms in the game and knowing which ones are still to come out, and their placement rules, is going to be a powerful predictor of who will be in a position to take advantage when they do. Strictly speaking memorising the composition of the tiles isn’t necessary to playing, but it is an important factor of playing well. For action tiles knowing what tiles are left to come out in phase two, three and four will be impactful on a player’s ability to plan ahead. As Mike Crabb pointed out to me, knowing the composition of the tile draws permits a player to put themselves in a position where they are better able to take advantage when they do come out- this is true for rooms and actions. Similarly, knowing the way that the turn structures evolve permits a player to modify their strategy so as to invest effort more effectively where it is important. To quote Mike directly:

Gold is the best example. I’ll gather gold and not use it knowing that it will give me an advantage in the last turn. if I can block the other player by taking the “make room” tile first, then take the “make room gold++” after, it’s a huge advantage and can swing the overall game”

Cave versus Cave doesn’t require much in the way of literacy – the titles of rooms are thematic rather than meaningful for what they do. However, it does require players internalise an awkward symbology where certain symbols have situational meaning. Whenever a dwarf symbol is used on a tile it means ‘Whatever count of dwarfs is active on the board’. The iconic language used is something you have to puzzle out rather than something that is usefully referential on its own merits. Good iconography is sufficiently clear that you don’t have to wonder.

Game flow is relatively consistent, but the number of actions a player can take during a round changes as do the options they have available. Mostly players are restricted to taking actions on their turn except for the ‘anytime action’ which converts certain resources into food. That can be done any number of times and at any time. What an action involves though can make the game seem like there is a malleability to turn order that isn’t there. You can’t use rooms in your cave unless you select an action that permits a ‘room action’, and in that case you get to activate one, two, three or four of your rooms. Since these have effects that act similarly to action tiles on the main board a player on occasion may take their one single action and what looks like four more. The game provides tokens to track when you’ve used a room action which helps alleviate some of the cognitive complexity of this. The tokens don’t help though with the ‘critical path analysis’ that would make the use of each action optimal in context.

Game state is moderately complex, but what’s more complex still is the puzzle of optimisation which is at the heart of play. Picking the right actions at the right time to give you the right rooms to accomplish the right tasks requires lots of things to line up and a fair amount of pre-planning to bring it all together. For example, to build a room requires you to choose that action. Having chosen the action, you need to have the necessary resources to purchase the room which requires you to have engaged in the necessary cultivation actions to generate them. The cultivation actions being available will depend on the room tiles you have previously bought. Let’s say though you can afford the room – then you need to place it. Placement requires an empty cavern square, which needs you to have excavated. Excavation needs you to take an action. Let’s say you’ve managed that though – you also need that empty space to have walls in a compatible configuration, and that requires you to have taken the necessary number of wall-building actions and placed those walls correctly.

That’s mostly what you’d expect of any game. The problem here is that all the planning done can be completely screwed over in a moment and perhaps with no obvious way to pivot. All of that has to line up in the right way at the right time, or someone else might well take the tile you have built your plan around and leave you unable to satisfy any other tile in the offering. To be fair, it’s rare that you’re left with no options but rooms are sufficiently powerful that you’ll almost certainly want the right one. Importantly this also needs a player to be reactive – to consider when new tiles enter the game and whether they’re better than anything else on offer. A good memory permits for planning ahead, to an extent, but everyone needs to be in a position to be able to react to unexpected changes in the flow of play.

All of this you have to do in a rhythm partially defined by your opponent since they will be eliminating options in a round as a consequence of taking them.

Scoring at least is straightforward, and while certain rooms work well together the game isn’t built on finding and executing upon synergies.

We don’t recommend Cave versus Cave in either of our categories of cognitive impairment. .

Emotional Accessibility

This is a cut-throat game – while you never directly impact on an opponent’s cave every single thing you do will whittle down their options. Some actions are so useful and powerful (excavation, furnishing and building walls in particular) that you’ll both constantly be trying to get them. They’re so heavily rationed too that almost certainly someone will find their plans absolutely flummoxed simply through unavailability of the necessary options. There are only eight rounds in the game and having one amount to nothing is frustrating. Two useless rounds in a row will put someone at an intense disadvantage and score disparities as a result can be significant.

Players with a compulsive need for competition may also find this an uncomfortable game to play, forcing players as it does to excavate when convenient and complete rooms only when possible. Often a cave is left looking like it was only half finished because things don’t often line up in exactly the way a player might like and it’s important to be pragmatic. Mistakes too can be very costly although wall placement at least has an in-game mechanism associated with the equivalent of an undo.

One of the weirder mechanisms in the game is the final, phase four action. If a player has more gold than their opponent they can place a wall and furnish a room in one single go. Now, it’s not always going to be the case that a player in the lead will have more gold than their opponent but it’s not a bad way to bet it. It’s almost the opposite of a catch-up action and you don’t need to spend gold to do it. Simply being richer than your opponent is sufficient to give you the bonus. This can only be done once, at the final round in the game, but it still strikes me as odd and an enabler of even greater score disparities than might otherwise have been possible.

All this said, the fact is that aside from both players drawing from the same limited pool there’s not a lot of player interaction. No way to undermine an opponent other than what is effectively ‘hate drafting’ – taking a thing you don’t want to deny it to an opponent. There is no way for players to gang up on each other because it supports a maximum of two. No upsetting themes, and no player elimination.

We’ll very tentatively recommend Cave versus Cave in this category.

Physical Accessibility

There isn’t any hidden game state here – everything is fully out in the open and thus no need to worry about card holders or secrecy. However, the player board is an awkward beast to manipulate. It’ll be covered in un-excavated rock to begin with, and gradually those tiles will make their way to the table for purchase. They’ll be replaced with new rooms that make up a cave, along with wall segments that are small and with huge importance for position. Likely a combination of all of these will be present at any one time.

A combination of things on the player board

The constraints on the player board are very tight, particularly when walls must be considered, but realistically there’s only a limited need for someone to play on the board. All the information on the board could be just as easily handled on a table or in a notebook. All it gives you is a place to track resources (and this is more easily handled externally, independent of physical accessibility needs) and a few cavern squares that provide resources when a space is excavated. The same experience could be had by keeping the board as a lookup and adopting a more free-form approach to play.

If that’s still not appropriate, Cave versus Cave does offer full support for play with verbalisation. Actions can be described unambiguously. A cavern’s various nooks and crannies can be likewise indicated with reference to cardinal directions and existing rooms. ‘North of the treasury’ or ‘east of the cave entrance’. ‘Place a wall in the north edge of the cavern south of my altar’.

We’ll recommend Cave versus Cave in this category.

Socioeconomic Accessibility

The manual makes use of the second person perspective and the box shows a male and female dwarf staring at each other across what is presumably a very tense neighbourhood fence. No objections here.

The back of the player boards

In terms of ‘cost per player’ this has a price of somewhere around £27 and supports only one or two people at a time. I’m not convinced it has the ongoing longevity to justify that price-point if cost is a factor – especially when you could get something like Patchwork from the same designer. While two player games are always going to be a difficult pitch for ‘value for money per player’ I’m not sure Cave versus Cave has enough in its box to warrant it even when taking that into account. For just a little more than the cost of Cave versus Cave you could buy Schotten-Totten and one of either Hanamikoji or Jaipur. Any one of them would be a better choice, and any two of them would undermine the case for Cave versus Cave in a heartbeat.

We’ll tentatively recommend Cave versus Cave here.


While the names of rooms are written on the tiles it’s not necessary to understand them. Cave versus Cave uses its own idiosyncratic symbology instead. There’s no formal need for communication during the game.

We’ll strongly recommend Cave versus Cave in this category.

Intersectional Accessibility

If memory impairments intersected with an emotional control disorder, I think the advantage conferred by knowing tile composition would have a significant impact here. A game that is already on the frustrating side would become increasingly so when an opponent possesses an asymmetric advantage.

Other than this we need to address the usual arithmetic of accessibility – a two player game means it’s statistically less likely that support will be available at the table if an issue requires a comparatively abled player. The chances of this though are likely to be known in advance.

Cave versus Cave plays reasonably quickly, but while it supports a solo mode it doesn’t permit seamless transition from one mode to the other if player count changes. You’re basically stuck with resetting the game for a new count. That said, it’s the loss of perhaps twenty minutes of play time and that doesn’t seem an unreasonable ask in comparison to other, heavier games.

One further issue to consider here is that the cut-throat nature of the competition does disincentivise players from offering spontaneous accessibility support – especially because pointing out an attractive option is almost certainly going to result in it becoming unavailable for another’s own use. As a result this may not be an appropriate game for intensely competitive pairings.


Sadly, this was Cave versus Cave’s second and final chance to earn itself an MLU recommendation – those games we can’t endorse on the basis of their fun often get a fair amount of love if they are open to the widest range of people. This alas was not to be.

Caverna: Cave vs Cave, Meeple Like Us, [CC-BY 4.0]
Colour BlindnessA
Visual AccessibilityD
Fluid IntelligenceD-
Physical AccessibilityB
Emotional AccessibilityC-
Socioeconomic AccessibilityC

Cave versus Cave is an odd blend of innovation and misstep. The two redundant forms of resource tracking are something I’d love to see in more games. And then they’re used in a tracker where there is no tactile differentiation. It recognises the last round is one to offer a differential advantage to make for a more interesting experience but then it hands that advantage to the player that is most likely already in the lead. It provides multiple sources for the same information on the board and then hides then away in unintuitive locations. It’s an odd duck.

We couldn’t muster much enthusiasm for Cave versus Cave, which is why it only got three stars in our review. It’s certainly a well-designed game but one where no matter how I try I can’t envisage being the one to bring it out of an evening. In order to pitch a game to someone you have to be able to sell it. ‘We should play this game because…’

I don’t have a ‘because’ for Cave versus Cave, and unfortunately this teardown didn’t do anything to change that.


A Disclaimer About Teardowns

Meeple Like Us is engaged in mapping out the accessibility landscape of tabletop games. Teardowns like this are data points. Games are not necessarily bad if they are scored poorly in any given section. They are not necessarily good if they score highly. The rating of a game in terms of its accessibility is not an indication as to its quality as a recreational product. These teardowns though however allow those with physical, cognitive and visual accessibility impairments to make an informed decision as to their ability to play.

Not all sections of this document will be relevant to every person. We consider matters of diversity, representation and inclusion to be important accessibility issues. If this offends you, then this will not be the blog for you. We will not debate with anyone whether these issues are worthy of discussion. You can check out our common response to common objections.

Teardowns are provided under a CC-BY 4.0 license. However, recommendation grades in teardowns are usually subjective and based primarily on heuristic analysis rather than embodied experience. No guarantee is made as to their correctness. Bear that in mind if adopting them.