Solo Play: Cosmic Balance, Not Good Luck

When I solo GURPS, or use an AI as a choose-your-own-adventure engine, the game can collapse in ways that have nothing to do with a bad character or a weak setting. Too many good things happen in a row. One clean win, then an ally, then a magic item, then another problem that solves itself. After a while I stop caring. The experience becomes a “so what.” I am just layering positive reinforcement on myself, and it feels stupid.

A randomized oracle does not save you from this. The dice are “fair” in the statistical sense and still allowed to give you a long string of exceptional results. Probability can, in theory, hand you a hundred great things in a row. That is legal. It will still kill the session. The same thing happens with an AI that has no instructions: it wants to be helpful, so it keeps saying yes.

I do not need the universe to be nicer or meaner. I need it to stop writing a story that has no opposite.

The needle

I keep a yin-and-yang score in my head. It is not a character ledger. It is a reading of my experience.

The easiest version, and the one I use when I just want the idea to be obvious:

• Ordinary good outcome: +5
• Ordinary complication or setback: –5
• Major wish-fulfillment (new ally, magic item, “I got everything I wanted”): +10
• A full combat encounter: –10

Combat is a heavy debit for two reasons. First, solo GURPS combat is expensive in time and attention. Second, it is one of the few places the campaign can actually end. Pricing that as –10 is honest.

If you want the needle to move more like an analog meter, roll the size of the event instead:

• Minor beat: 1d6
• Average beat: 2d6
• Major beat: 3d6

Apply the roll as a plus or a minus, depending on which way the event actually felt. A small favor is +1d6. A small snag is –1d6. Finding a real ally or a magic item is +3d6. A full fight is –3d6.

I used ±5 and ±10 in this article because they make the theory of cosmic balance easiest to see. If you already think in 3d6, roll it. The dice give you variance: a “major good” might only be +8, or it might be +16, and that changes how overdue the next pull feels.

The number does not have to snap back to zero after every scene. Think of an analog meter, not a magnet. The needle jumps. It can sit at +10 or –10 for a while, and the game can still feel interesting. When the next real choice comes up, I glance at it. Is a rebalance overdue? Which way should this beat lean?

Asymmetry

The system is not trying to keep life fair. It is trying to stop the good side from becoming weightless.

A string of bad luck is still a story. Two fights in a row, things getting worse, a stretch you have to climb out of: that happens in adventures. If I am already at –10, another –5, another –3d6, or another combat is allowed. The needle can slam into the red. That can feel like a challenge.

A string of big pluses is the thing that drains my interest. I am not forbidding two great things in a row. I am wary of them. The second gift is usually what starts the feedback loop.

The needle is important, not absolute

If the score is sitting at –40 and the characters find a safe town, take a room at an inn, and spend a week healing, that can be a +10, a +3d6, a +20, or a full reset. If the fiction supports a pause, I do not argue with the meter. A recovery after a justified grind is not the same as another unearned win.

What I am actually guarding against is the consecutive, cheap gift streak: win, then ally, then item, then another clean victory, until nothing is at stake. A week of rest after a brutal stretch is the opposite of that. The inn is interesting because the needle was already in the red.

Three rules of thumb:

1. Watch the needle when the next beat could go either way.
2. Be more wary of another +10 (or another +3d6) than of another –10.
3. If the story itself offers a genuine pause, rest, safety, time to heal, you can swing the needle hard or zero it out. Do not invent that pause just to make the number look nicer.

The meter tells you when a rebalance is overdue. The story tells you whether this is the moment to do it.

What I tell the AI

A dice oracle can theoretically give you a hundred great things in a row. An uninstructed AI will usually do the same thing by being agreeable. The balance score is what forbids both.

I do not tell the model to return the score to zero every scene. I tell it to track the score for my experience, not the character’s success, and to consult it when the next significant choice appears.

Something in this shape:

Keep a running Balance score for the player’s experience, not the character’s win/loss record. Use either the simple values or dice: Ordinary good: +5 or +2d6. Ordinary complication: –5 or –2d6. Minor beats: ±1d6. Major wish-fulfillment (ally, powerful item, clean “I got everything I wanted”): +10 or +3d6. A full combat encounter: –10 or –3d6. The score does not have to return to zero after every scene. It can sit high or low for a while. When the next meaningful beat could reasonably go either way, use the current score as a bias. Be more wary of stacking major pluses than of stacking penalties. A second combat or another stretch of bad luck is allowed; it can mean times are getting tough. A second unearned gift in a row is the thing to avoid. If the characters actually stop, find safety, and recover, you may swing the score hard or reset it. Do not invent that recovery just to tidy the number.

That is the whole trick. Cosmic balance here is not karma. It is a governor. It protects the game from luck or positive bias - especially from good luck and wish-fulfillment.