The Resolution Fragility Score: When the Market Itself Is the Risk
Introducing RFS and what an 18-day watch on a single prediction market revealed about why some markets look liquid right up until they break.
TL;DR: Every prediction market is really asking two questions: what is the probability this resolves yes? and can this market actually be resolved cleanly? Only the first one is priced into the market. The Resolution Fragility Score (RFS) is ArAIstotle's answer to the second — a 0–100 number, calculated independently of the probability estimate, that measures how vulnerable a market's settlement is to being reversed, delayed, or fought after the fact.
Higher scores mean the outcome could be more fragile, even if trading looks liquid today.
For the last 18 days, the Truth Terminal has been quietly tracking a single prediction market.
The market is priced on a specific outcome involving MicroStrategy's Bitcoin position. Tens of thousands of dollars have moved through it. Traders have come and gone. The crowd's price has shifted.
And the entire time, ArAIstotle has been telling anyone watching that the market is structurally fragile, not because the probability is hard to estimate, but because the settlement mechanism itself has design choices that make a clean resolution unlikely.
ArAIstotle's signal for this is the Resolution Fragility Score, or RFS. It is the layer of the Truth Terminal we have not talked about enough, and the layer that we think the next generation of prediction market traders and trading agents will need most.
Every prediction market is asking two questions
Every prediction market is really asking two questions, even though most platforms only price one.
The first is the obvious one: what is the probability that this resolves yes? The market price is the crowd's answer. It is what most analysis, most coverage, and most trader attention focuses on.
The second is the structural one: can this market actually be resolved cleanly? It is rarely asked out loud, and almost never answered with a number. But it matters just as much. A market with a 70% probability and a clean resolution mechanism is a fundamentally different instrument than a market with the same probability and a contested one. One pays out predictably. The other turns into an argument months after the bell.
The Resolution Fragility Score is ArAIstotle's answer to the second question.
What Resolution Fragility Score actually measures
The official definition, taken directly from the Truth Terminal UI:
Resolution fragility score (RFS), 0–100: how vulnerable a settlement is to being reversed, delayed, or fought after the fact. Higher scores mean the outcome could be more fragile, even if trading looks liquid today.
That last clause is the philosophical heart of the feature. Even if trading looks liquid today. Liquidity is not a measure of resolution quality. A market can attract real volume, tight spreads, and confident-looking pricing — and still be sitting on a settlement mechanism that breaks the moment the resolution date arrives.
RFS scores are interpreted in ranges. The MSTR market we have been tracking has been sitting at 80/100 — HIGH for the duration of the 18-day watch. The reasoning has updated as conditions shifted, but the structural assessment has not.
A HIGH RFS reading typically flags some combination of:
Single-source reporting from a private entity. When a market resolves based on what a specific company publishes about itself, that company's reporting schedule, framing, and accuracy become resolution risks.
Fast-moving underlying conditions. Markets tied to assets in periods of high volatility — cryptocurrency prices, breaking geopolitical events — face a narrower window between data availability and resolution.
Ambiguous resolution language. Words like exceed, by, confirmed, and officially have all been contested on prediction market platforms before. Markets with imprecise criteria carry built-in resolution risk.
Limited or politicized data sources. Markets that depend on a single oracle, a single news outlet, or contested data feeds are structurally exposed to dispute.
Active misinformation potential. Markets on topics with viral misinformation cycles can be moved by false claims faster than verification can correct them.
The point of RFS is not that high-fragility markets shouldn't be traded. It is that they shouldn't be traded the same way as clean ones.
What an 18-day watch looks like
The MSTR-Bitcoin market is a textbook high-RFS case.
The market depends on data reported by MicroStrategy itself, a private entity that publishes on its own schedule, with its own framing. It depends on Bitcoin's price during one of the most turbulent stretches the asset has seen this year. It depends on external reporting that often lags the underlying events. And it depends on the resolution criteria, with language that has been contested on prediction market platforms before.
ArAIstotle has been generating an RFS reading on this market every time the Terminal has refreshed for the past 18 days. The score has stayed HIGH the entire time. The detailed reasoning has updated the underlying sources change as Bitcoin moves and as MSTR's reporting evolves, but the structural conclusion has not.
Anyone trading this market on the Truth Terminal has been able to see, in real time, exactly which fragility factors were active and why. Not as a hidden internal model. As a public number with the reasoning attached.
This is what continuous verification of market quality looks like in practice. Not a one-time audit. A live read on whether the market is the kind of instrument that can be cleanly traded on, updated as the conditions shift.
Why this matters for prediction market traders
If you trade prediction markets, this is the part that should land.
A probability estimate tells you whether to take a position. An RFS reading tells you how to size it, how long to hold it, and what to watch for between entry and resolution.
A market with a 60% ArAIstotle estimate and a LOW RFS is one kind of trade: high-confidence, clean settlement, hold to resolution. A market with the same 60% estimate and an RFS of 80 is a different trade entirely: smaller position, faster exit, watch for resolution
drama, be ready for ambiguity.
The crowd does not distinguish between these. The market price does not distinguish between these. ArAIstotle does and now exposes that distinction publicly on the Terminal.
This is the literacy layer that has existed in equity markets for decades. Equity traders do not just look at price. They look at liquidity depth, spread quality, regulatory risk, and a dozen structural factors that determine whether a position is cleanly tradeable. RFS is the prediction market equivalent.
Why this matters for AI agents
This is the part that matters for the agentic economy.
Autonomous AI agents trading prediction markets cannot make these distinctions on their own. They see a probability and a price. They do not see the structural vulnerabilities of the resolution mechanism. They do not weigh single-source reporting risk against multi-source consensus. They do not flag when a resolution criterion contains ambiguous language that has been contested before.
An RFS reading, exposed as a machine-readable signal alongside the probability estimate, gives agents what they cannot derive themselves: a structural risk overlay on every market, updated in real time, sourced and auditable.
This is the missing piece for a responsible agent trading on prediction markets. Not a better probability estimation. Better market triage. Telling the agent not just what to think about the outcome, but whether the market itself is the kind of instrument that can be cleanly traded on.
Where RFS sits in the verification-native stack
Readers of our last post on the verification-native prediction market will recognize this as the structural layer underneath the probability deltas we have been writing about. Two distinct signals, doing different jobs:
The delta tells you the crowd is mispricing the outcome.
The RFS tells you whether the market itself can be cleanly resolved on.
A 40-point delta on a low-RFS market is a trade. A 40-point delta on an 80-RFS market is often a trap. The trade looks the same on the surface. The actual risk profile is completely different.
Both signals matter. Neither is complete without the other. Together, they are what we mean when we say verification has to be built in, not bolted on.
The future of prediction markets is verifiable price discovery
The future of prediction markets is not price discovery alone. It is verifiable price discovery at every layer, including the layer that asks whether the market itself is structurally sound enough to discover anything on.
ArAIstotle is building that verification stack today. The Resolution Fragility Score is one of its quieter components, and one of its most important.
Try the Truth Terminal to see RFS running on every major market: araistotle.facticity.ai/terminal/market.
Trading a market with high-resolution fragility? Curious what RFS reads on a specific question? Reach us at [email protected].
