Game Chef 2016: Playtesting Sands Robot

We had a blatant omission to our submission for Game Chef 2016: Playtesting. Technically, Game Chef 2016 design is over. We’ll need to leave the rules and released materials alone for now while the reviews and judging are going. Still, we’ve got something worth trying so we wanted to give it a shot. The short story is: “Oops, it’s a little broken.” The longer story (below) is that it seems like there’s a fun game in there, with the right tweaks. We’ve got some suggestions if you end up trying to play it as it is today. Once the judging is over, we’ll release an update.

Glaring Errors

Printing went fine. It looks like the PDF makes for a good (if not quite standard) card and board size. Right off the bat after setup we noticed that the robot pieces have no distinguishing marks. We had to mark them with numbers to tell the robots apart.

When picking the first human, the instructions call for a dice roll. Not that most people don’t have ready access to dice, but they aren’t used anywhere else in the game. We opted to shuffle up the role cards and draw for first human (then reshuffle to distribute robot roles).

The instructions have the human draw cards before removing repair cards and shuffling the deck. Obvious one to figure out, but amidst the other confusing aspects of the instructions, this was a few more minutes of chaos getting going.

Clearly the kind of mundane things that should have been caught ahead of time! So it goes.

The First Game

We played with four players (fairly standard choice). The first game was with all cards visible (no hiding hands). We didn’t play it all the way to completion, because of the biggest hole we found: nothing really got built! On top of that, it took forever for the first two transfers (triggering repair activation) to occur. Clearly the robot move balance was way off.

Most of what we found were clarification issues. I’ve moved those to the bottom of this post

About halfway through we decided we understood the rules enough, and were getting annoyed with how little was actually happening. Time to play right, with some fixing up.

New Game, New Rules

The lack of real action made a few adjustments necessary:

  1. We changed the starting positions to allow the human to place all markers (robot and human). We also made a note that a good strategy is to start some robots near each other.
  2. We removed Construct cards entirely, and treated all Construct cards as malfunction cards. Construction now occurred when a robot moved (or was moved) to a space without a reflector.

Both are crucial to making the game a little more lively.

The second game played fairly well. We had one major snag- we had the hardest time keeping track of which robot went next. In the end, my son lent us a hand in the form of a turn marker:

Its the pacifier

Sands Robot playtesting turn marker

Beyond that, gameplay went pretty well. We robots managed to get enough transfers going early to hide any obvious conclusions of who was who. The human for the round got a lucky break when a second repair against me failed after a first succeeded, ruling me out entirely. She ended up correctly identifying the sentient robot about 20 minutes in. Not a bad little round.

Still, it was unanimous that the reflectors weren’t doing anything. Even with the change of building on move, less than a quarter of the board ended up built on. We made a drastic change for the next game. But first…


It was quite challenging to keep information hidden. An idea I abandoned early, privacy screens, nearly came back into the picture. We found a solution to the problems that arose, although more likely exist. So far, the game seems to rely on players willing to stick to the spirit of the game to work. Regardless, here’s what we learned:

  • It’s important to draw the Robot Action card into your hand, find your piece on the board, determine the result of your move, and finally announce your move and discard face down. This needs to be done regardless of whether you are currently sentient or not. Without doing this, the tendency was to easily give away your status by, e.g., announcing your move and then having to find your piece, making it clear you didn’t have to think about which move you were going to make
  • When Ctrl+Alt+Del is played, all robots should give their interface cards to a single robot, who then distributes them. We had a brief look around the robots trying to figure out how to even the hands without giving them away, and this was the solution


We had time for one more game. Building was doing essentially nothing; neither was moving. So we confined quarters a little. The board got stripped way down to a 3×3 square. We kept the same rules about building on move, so now it was a little more likely that the board might get filled.

It was also clear that there wasn’t quite enough pressure not to waste repair cards. We removed one from the 4 player game, down to 5 repair cards for three robots. This turned up the pressure quite a bit, since it removed quite a few simple logical tests to find the sentient robot. As a consolation, we did away with the multiple-repair rule and allowed back-to-back repairs against the same robot. After all, its a pretty big gamble to go two repairs deep into one robot unless you’re already sure.

This game played really well. I was the human. I found myself trying to track more information than I could to determine who the sentient robot was. In the end, all I could remember was which transfers had occurred and a few File Browsers I had done early on. My strategy was to herd the robots into the transfers I wanted to test. I got a lucky break when I was able to pin one robot for four turns and keep the other two interacting and giving away information.

Towards the end of the game, once it was clear I was close to figuring out the sentient robot, the tension of the game let up a little early. The robots began malfunctioning a lot just to sabotage the possibility that I could wait out the factory build. I think the next play will add this as a mechanic; once the repair cards come into play, the robots can win by destroying all solar collectors.

Instruction Clarifications

Below are the remaining clarifications that were discovered:

  • Two robots being the target of an interface, not two interfaces, triggers the repair card addition.
  • The repair card is supposed to be removed from play regardless of whether the repair succeeds or not.
  • The interfacing rules are completely confusing. Nobody could figure it out until it was explained verbally. First, the terms needed to be cleared. We decided on “initiating robot” and “target robot”. The initiating robot gives one interface card to the target robot, who adds it to their hand. The initiating robot never changes their status, regardless of the interface that occurs.
  • If a robot lands on a space with more than one other robot, the robot moving into the space (and initiating the transfer) chooses the target.
  • Clearly the robots could use some reference cards to see what moves are available. And if you could tell the robots apart, the cards could even match the robots!
  • Summoning a robot triggers an interface, but does not trigger construction
  • No diagonal movement is allowed
  • Skipping a robot’s turn still results in two turns before the human makes their play. Its useful to put the CPU Overload card in front of the robot to remind about the skip, then send it to discard once the skip occurs

Leave a Reply

Your email address will not be published. Required fields are marked *