Ollie Ollie Oxen Free

(This review originally appeared as a blog post of mine during IFComp 2013.)

Ollie Ollie Oxen Free is a primary school-based adventure of rigourous puzzling in which you play a teacher who must rescue a series of trapped students in the wake of some kind of bombing. The source of the threat isn't specified, or ultimately important, at least as far into the game as I reached before giving up, which I did after 145 minutes.

Ollie's ambitious design supports all of the students independently. You can talk to them, order them about separately and have them act as the instruments of puzzle solving for you, which is necessary because the attack has left you too weak to perform any dexterity-demanding tasks. To successfully marshal them to help you help them rescue each other is the kind of feat which will convince you that you could organise a team of green berets. But with great mechanics must come greater implementation. The tools the game gives the player to do what is being asked of them are underpowered, and there are a lot of bugs and oversights. Also, I don't consider it acceptable to have a parser game say things like: "If that command didn't work, please enter it again," or "It looks like you've completed that part of the walkthrough, but I'm not sure." My guess is the author ran out of development time before IFComp.

Detailed discussion with spoilers ahead:

The layout and presentation of the school building has a realistic logic and a pleasing adventure game aesthetic in terms of the distribution of remarkable features. The descriptions depict a school environment for little kids through an adult's eyes. The teacher's observations on the naff posters and simplistic kiddie artworks express light cynicism, but his subsequent earnest interactions with the kids show how he can compartmentalise adult thoughts.

The game is good at introducing new gameplay mechanics, sometimes through cueing in the prose and sometimes through explicit help messages. And there are a lot of mechanics: SHOUTing to locate kids, THINKing about people or topics, ASKing kids about people or topics, and ordering kids to perform actions. Kids can be spoken to from up to a room away, made to follow you around, or to collect and use various props. They also have different personalities and fears that you need to manage, and these are a source of cute and touching observations of the kids' personalities, as well as a source of puzzles.

The interplay of all of these elements is particularly complex in light of the game's microscopic-leaning scale. The children don't react to broad commands, only to specific ones like SAMIR, GO WEST. ASHLEY, PUSH THE MAT NORTH. TYRONE, GET THE YARDSTICK. In turn, you are limited in being able to have only two children follow you at any particular time, and that each of those children can only carry realistic amounts of equipment.

I am not of the school of players who universally reject inventory limits. In terms of generating interesting logistical challenges, I think Ollie's limits are clever ones, but the trouble for this game is that the number of commands required to try out even a moderately novel puzzle solution can be huge. You need to muster the right children in the right locations, have them carrying the right things, then find the right commands. If your idea doesn't work out, it will probably take at least twice as many commands to undo everything that has been done and to redo it in a slightly different way. The problems of logistical optimisation currently comprise the game's major challenge. And again, I don't oppose this per se. Such challenges can be satisfying to solve, leading the player to a deep engagement with the gameworld. But the player has to be able to have great faith in the reliability of the game's feedback if they're not to feel that they're in danger of wasting their time. Ollie did not generate that faith in me.

I hit all kinds of bugs and problems during play. The prose made incorrect assumptions about what knowledge I had acquired so far, characters spoke out of turn or from out of earshot, crucial conversation topics didn't register, vital items weren't mentioned in room descriptions, mid-puzzle feedback failed to suggest I was making progress.

Bugs and oversights can be fixed, though in the meantime the game much harder than need be and a frustrating vision of what it could be. I think the trickier issue lies in the realm of speculation. Inform has technology in place that would allow Ollie to dispense with a lot of its micromanagement. I can imagine a version of the game in which children can be told to go to rooms, or to collect a particular item and return, etc., with single commands. I'm sure this would be extremely challenging to program, but I believe it could be done, and would eliminate all of the time and hard slog currently involved in trying to execute ideas which aren't necessarily complicated, but which require tons of commands and perhaps gritted teeth to even broach. The result would be a different game – not massively different, in fact the core design would remain the same – but that game would not present the extreme optimisation problems the current one does. Atop it all, the current game admits to the player that it doesn't understand its own state.