Monday, October 22, 2012

Thoughts on Game Design and Vision

A few players have recently expressed some frustrations with NEO Scavenger, which brought up some questions about core design. In particular, the concerns centered around the combination of permadeath, save games, and randomness.

Game Design

The frustrations stem from the game ending, irrevocably, as a result of chance in-game hardships. I think most players would agree that investing significant time into an RPG character, and then losing it to a die roll, is a real buzz-kill. Some will sigh and re-roll a character, others will walk away for a while, and still others may rage quit, never to return. Whichever is the case, such an occurrence in-game is poor game design. Instead of the player suffering a penalty due to poor choices, they suffered due to bad luck.

A powerful storytelling tool. Also, sucks when it happens to you.
In defining NEO Scavenger, I eventually settled on the idea that I was making a single player computer RPG that could simulate many of the experiences had in a pen and paper RPG session. And many of the game's systems are successes in that regard. For example:

  1. players can customize their character
  2. there are multiple ways to use the player's skills and items meaningfully
  3. there are elements of strategy and tactics
  4. there are setting and plot elements to discover
  5. there are options to role-play moral choices. 
With the addition of a permadeath save feature, it's even possible to save progress, walk away, and return later: just like one does in a role playing session with friends.

One thing NEO Scavenger lacks, however, is a game master (GM). A good GM does many things in a session. From creating the story, to arbitrating the outcome of conflict and struggle, the GM is a human element which ensures the game is fun for the players. And when an unlucky die roll threatens to spoil the game, a good GM can find ways to fudge the outcome, or interpret the roll in a way that adds drama, rather than removing the fun.

Unfortunately, when misfortune strikes in NEO Scavenger, there's no such GM to ensure the player is having fun. And quite often, this misfortune is game-ending. So, what can be done to fix this?

Save Games

On a few occasions, players have suggested save games as a way to mitigate this problem. With checkpoints and save slots a common feature in many games these days, I can see why. It's an easy way to ensure a player's invested time is not lost. Like an insurance policy, the player can resume where their last save left off, minimizing the amount of time spent replaying existing content.

However, as I've discussed before, knowing one can just reload the game at any time robs the game of two powerful tools: fear and consequence. Thankfully, most players seem to be on board with this philosophy. They enjoy the game exactly for this hardcore nature, and diminishing that would diminish the game as a whole for them.

(Incidentally, players who want to can still circumvent the save-delete by making a copy of the save file, a la Nethack. I'm happy to leave this intact for players who really want it.)

Ultimately, I want the player to have the convenience of playing when they want, but I don't want to use save games to fix what may be a design flaw in the mechanics. So if save games are just a band-aid, what's the actual problem?


From the outset, NEO Scavenger has leaned pretty hard on randomness. From random map hexes, to random combat outcomes, Math.random() was a crutch I used liberally.

Randomness isn't inherently bad, mind you. It's a pretty valuable tool for making the game interesting more than once. It can provide fresh environments to explore, a seemingly deeper AI, or random bouts of luck or misfortune to tip the scales of drama.

However, too much randomness can be bad, especially where life and death of a player's character is concerned. If a player feels like their choices don't matter, and that the game will simply reward/punish them at random, then they won't enjoy the game. Or, at least, the type of player I'm building this game for (i.e. me) won't.

Note to self: These are decision-making tools, not decision makers.
One of NEO Scavenger's earliest failures was combat. There was a random chance of injury imposed on the player for participating. In many cases, that injury was instantly fatal.

Players reported feeling that the game would randomly kill them, and felt frustrated because there were no mitigating strategies. One simply hoped for a good outcome each turn. In fact, most simply chose to ignore and avoid combat entirely, instead focusing on more rewarding and enjoyable areas of the game.

Fortunately, we were able to identify this, and we found ways to make combat more interactive. We gave the player more verbs in combat, and more nouns for them to act upon.

Instead of a short "overall injury" meter, the player now had multiple wound locations, blood supply, pain, and infections to manage. Instead of deciding whether to "hit" or "run," the combatants now had additional moves they could choose from, like "trip," "tackle," or "lure into trap." And the addition of situational factors, such as terrain, morale, cover, lighting, and stance, meant that the player could opt for different strategies as the battle changed.

What this did was allow players to see trouble in its early stages, and gave them time to deal with it. There was still randomness, but it was relegated to minor victories and losses. The overall life and death of the player was largely in their hands, and they were left to choose when a reward was worth the sacrifice.

This is still a weak point for NEO Scavenger, and effort still goes into finding ways to strengthen it. I think we're on to something, though. In the linked example above, I think we've identified some tangible causes of frustration, and we're discussing in-game solutions that might help. It has a familiar ring to it: games are more fun if the player's choices determine victory or defeat, not random chance.


While randomness is surely a weak point for NEO Scavenger, there's one other area I'm keeping an eye on. That area is the replayability, especially in the early game.

If a player is afraid of dying, that's a good thing, in my mind. It puts them in the right frame of mind for enjoying this game of survival, and making each decision count. It's what most players cite as their reason for enjoying NEO Scavenger, and it enhances the thrill of the game.

However, some players have reported feeling turned off when dying, to the point that they do not want to play again. This might point to a separate problem, one which has more to do with the fun factor in the early portion of the game. In other words, it seems that if the early game was lots of fun, players probably wouldn't mind dying as much.

For example, I'm always willing to start a new game of Civ 4/5 or Galactic Civ II. The early game is one of my favorite parts. There are even times when I'll cut an existing game short so that I can restart.

Mind you, I still enjoy the mid-late games too. But I never feel disdain as my civilization starts to collapse, because I look forward to the next go.

In fact, that was often the case for me with pen and paper RPGs, too. Rolling new characters was "half the fun" for me. I loved exploring other strategies, backstories, and those first few struggles to get one's feet.

I'd like that to be the case for NEO Scavenger, too. So when players say that the Game Over screen gives them pause, I start to wonder if NEO Scavenger's mid-late game is the compelling part, and the start is just a barrier. Is there something really fun happening just before death, that the player is trying to get to as quickly as possible next time? Is the early game too boring or tedious?

So far, when I've asked about this, the general response is that the early game is ok. And frankly, the game is probably short enough that there's little to no mid-late game anyway. Still, it's on my mind. More something I'm monitoring than actively working on.

Perhaps addressing the randomness issue will have an effect on the early game? If players feel more enabled, and that their choice caused death, they'll be more engaged in the character building process? Or perhaps the character building process itself is too limited? Perhaps there's not enough encouragement to experiment?

One thing at a time, I guess. Changing too much at once can muddle the results, so we'll deal with the obvious problem first. Perhaps the way forward will be more obvious after this obstacle is cleared.