ogsfrompoly

Methodology

How ogsfrompoly tests Polymarket wallets for repeatable skill, scores retrospective alerts, and decides what is safe to publish.

ogsfrompoly asks a narrow question:

Which Polymarket wallets look skilled after luck is stripped out, and do their moves keep working when we score them later?

The answer is published as a retrospective scorecard. We do not publish live trade ideas, open-position details, or wallet leaderboards.

Current scope

The public track record currently covers two Polymarket areas:

This narrower scope is deliberate. The goal is not broad category coverage. The goal is a clean audit trail: skilled wallets, retrospective alerts, and published outcomes that a reader can check later.

How a wallet qualifies

A wallet first needs enough resolved history to test. Here, an event means one real-world question being settled on Polymarket, even if that question has multiple markets, outcomes, fills, or token legs. The minimum is at least 20 events that have resolved within the measurement period.

Only resolved markets count. Open positions are ignored. For each eligible wallet, we build an only-realized PnL series: one profit-and-loss value per resolved real-world event.

We group at the event level because one real-world prediction can involve multiple fills or market legs. Counting each fill as a separate decision would make one position look like many independent bets. Counting at the event level keeps the test closer to the thing we care about: whether the wallet repeatedly picked the right side of resolved Polymarket events.

The skill test

Skill is measured with an event-level sign-randomization test.

For each eligible wallet, the inputs are:

Each simulation keeps the size of every event result but randomizes the sign. For simulation j, each event gets a random sign:

s_{j,i} in {-1, +1}

The simulated no-skill total is:

T_j = sum(s_{j,i} * abs(x_i))

The p-value is:

p = count(T_j >= A) / B

In words, the p-value answers:

"If this wallet had no skill, how often would random sign flips produce a total PnL at least as good as the wallet's actual total?"

A wallet is treated as skilled only when that p-value clears the pre-registered threshold: p <= 0.05. A large raw PnL number is not enough.

Controls before publication

The public scorecard should measure repeatable trader signal, not address management or market microstructure artifacts.

Before a wallet can affect the published record, the pipeline applies controls for coordinated multi-wallet clusters, concentrated counterparty patterns, and wash-like behavior. These controls either remove or demote signals that can make one address look more skilled than it really is.

How alerts are scored

Weekly statements score the period after the fact. They report:

Monthly statements add the business side of the project: revenue, operating expense, net result, and runway.

The published categories are politics and macro-finance. Anything outside the current public scope is dropped from the landing-page statement surface instead of being merged into another category.

Every statement is retrospective. The site publishes whether historical signals worked. It does not publish instructions for trades in flight.

Academic foundation

The methodology is grounded in two 2026 papers on prediction-market skill, trader persistence, and market outcomes:

Both papers point to a practical lesson: short-term realized PnL is a weak proxy for skill. A wallet can get lucky, and a lucky wallet can stop winning.

That is why ogsfrompoly starts with a statistical skill test, then publishes whether the resulting signals actually worked in the public track record.

Disclosure policy

The site is open about methodology and aggregate results. It is closed about anything that could front-run a position or expose an individual trader's identity.

These rules apply to the rendered HTML page, the raw .md feed, RSS, and llms.txt.

When a published number sits close to the line, we delay by 30+ days or anonymize the category before publishing. When in doubt, we don't publish.

What this is not

ogsfrompoly is not an investment recommendation, trading signal service, or claim that any wallet will keep winning.

It is an audit trail for one specific method: find wallets whose resolved history looks different from a sign-randomized null, follow their retrospective alerts in the categories we publish, and show the results.