How the DraftGamers Index works.

01What the scores mean

Every rated competitor carries three numbers.

• SKILL measures demonstrated competitive standing: results, consistency, level of opposition and longevity.
• HYPE measures cultural weight: attention, reach, influence and share of the conversation.
• Draft Score is the composite. It is not an average. HYPE and SKILL carry different weights by category, because a chess grandmaster and a variety streamer are not doing the same job.

All three run 0–300 on the same scale across every category. A 240 in poker and a 240 in Counter-Strike mean the same thing.

The bands

Band names are the ones shown on the live product. Counts are the current published edition.

The gap

The distance between a person's SKILL and HYPE is a rating in itself. A large positive gap indicates underpriced ability and is flagged Sleeper. A large negative gap indicates attention running ahead of results and is flagged Overvalued. Everything between the two is Fair. Both directions are published, including where the flag runs against the top of the board.

02How the scores are built

Inputs

Each rated competitor is tracked across a structured record covering six families of evidence:

• Competitive results. Placements, titles, level of opposition, career earnings, years at level.
• Consistency and form. Recent form, performance variance, floor as well as ceiling.
• Longevity. Time at level, career arc, peak standing, durability.
• Audience. Reach and engagement across platforms: followers, average and peak concurrents, hours watched and views.
• Cultural signal. Conversation share, moment creation, search interest, influence on the field.
• Trajectory. Direction and rate of change: growth, movement and momentum.

The composite

Draft Score is a per-category linear composite of HYPE and SKILL. These are the current weights, published in full:

HYPE outweighs SKILL by roughly two to one in every category. This is deliberate and it defines what the index measures. Draft Score rates standing in competitive gaming culture, which is a function of attention as much as achievement. It is not a measure of competitive ability alone. SKILL is published as a separate column for readers who require competitive performance on its own.

Normalization and cadence

• Normalization. Categories are scored internally first, then normalized to the shared 0–300 scale. This is what makes cross-category comparison valid.
• Refresh. Scores are recomputed on a weekly cycle. Editions are locked at close. A score inside a locked edition does not change afterwards except through a published correction.
• Human review. The index is not fully automatic. Data feeds the model and a review pass catches outliers, disputed results and events the inputs cannot see.

What we publish and what we hold

We publish the scale, the bands, the input families, the per-category weights above, the cadence, and the reasoning. What we do not publish is the dimension-level rubric that produces the HYPE and SKILL inputs themselves.

This is deliberate. An index whose full scoring rubric is public becomes a target to optimize against, and an index that can be farmed stops measuring anything. The same reasoning applies at every major credit and consumer rating agency.

03Coverage and inclusion

1,526 competitors are rated across six categories: Creators, Esports, Strategy (chess and strategy games), Gambling (poker and skill gambling), Speedrun, and Crypto.

Who qualifies

• Individuals only. A row is a person with a name and a public record. Organisations, teams, studios, publishers, games, tokens, protocols, platforms and guilds are excluded. Founders and executives are eligible; the companies they run are not.
• Measurable presence. A person enters the universe by demonstrating competitive or cultural presence in one of the six categories at a level that generates trackable data. Coverage is not a judgment of worth. It is a judgment of whether the person can be rated honestly with the data available.
• Nobody buys their way in. Inclusion has never been for sale and will not be.
• Inactivity. Competitors who go inactive keep their scores and are flagged rather than deleted. The historical record is retained in full.

Known limitations

These are the current weak points in the data and the priorities for the next version. They are disclosed rather than omitted:

• SKILL measures current standing, not peak. A retired champion scores below an active one, which overwrites peak achievement. Splitting current and peak SKILL is the top priority for the next version.
• Creator SKILL is rubric-derived. Creators have no competitive record, so SKILL there is estimated from production scale, longevity, earnings and competitive-adjacent output. It is an editorial estimate, not a measurement, and it is not comparable to a match record.
• Esports discipline tagging is incomplete. A large share of esports rows carry no discipline tag yet, which limits game-level cuts.
• Crypto inputs are thinner. Crypto HYPE is estimated primarily from follower counts; better reputation inputs are identified and not yet integrated.
• Recently added rows are provisional. New entries carry a low audience-confidence flag until verified against platform data.
• Net-worth and earnings figures are estimates unless drawn from an exact public record. They are published as ranges rather than as precise figures.

Requesting review

If you believe someone is missing, tell us at support@draftgamers.com. People are added because they qualify, not on request. Submissions are reviewed regardless.

04Independence

DraftGamers sells media, campaigns and data. DraftGamers also rates people and organisations who may be involved in those transactions. That conflict is real. It is the same conflict every rating agency carries, and it is managed by structure rather than by assurance.

• No commercial relationship influences any score. No sponsor, advertiser, partner or client can move a number in either direction by any means.
• Nobody on the commercial side has score authority. Score decisions sit with the index function and do not route through anyone who sells.
• Rated parties get no editorial input. No preview of a score change, no approval, no negotiation.
• Relationships are disclosed where a rated competitor or organisation has a commercial relationship with DraftGamers, at the point the rating appears.
• No advance notice. Scores publish to everyone at the same time.

Roster conflicts, where a sponsor's creators are rated, are routine rather than exceptional. They are handled by disclosure and by the first rule above, not by removing anyone from the board.

An index that can be influenced holds no value for any party, including the one seeking to influence it. Failures of this policy should carry a commercial cost to DraftGamers, and are treated accordingly.

05Corrections

We will get things wrong. When we do, the correction is public, dated and permanent.

• What triggers a correction: bad input data, a misapplied rule, a result recorded incorrectly, an identity error.
• What does not: disagreement with a score. A rating is an opinion supported by a method. Disagreement with the outcome is not an error.

Process. Report an issue to support@draftgamers.com. Verified errors are corrected at the next close, or immediately if material. Every correction is logged here with the date, what was wrong and what changed.

Corrections log

06Accuracy record

The index makes public calls. This is the record of how they resolved.

Every prediction is timestamped before the outcome, locked, and settled against the result. Where a betting or prediction market priced the same event, that price is shown alongside. Incorrect calls are published with the correct ones. An index that reports only its successes is not a record.

07Citing and licensing

Free to cite. No permission required, no fee, no form.

Required attribution:

You may

• Quote scores and rankings in reporting, research, decks and analysis.
• Reference the index by name.
• Link to any page on this site.

You need a licence to

• Redistribute the dataset.
• Build a product or feature on top of Draft Scores.
• Use the index in a commercial transaction.
• Access historical editions or the full rulebook.

Licensing and diligence requests: partners@draftgamers.com.

08Version history

The DraftGamers Index is published by DraftGamers. Editions are locked at close and archived permanently.