The GSI Method — How Poker Tournament Structures Are Measured

Why poker tournament structure matters

A tournament may be described as deep, fast or player-friendly, but those descriptions are often subjective. GSI replaces impression with a consistent measurement framework focused on the published structure itself.

The objective is not to measure the event as a whole. It is to make tournament structure measurable, readable and consistently measured.

What GSI analyzes

Each eligible structure is analyzed in depth by the GSI Intelligence Engine, which is frozen for the season.

Only published schedule data enters the calculation: starting stack, blinds, antes, level durations, players per table and late registration close. From these, the engine derives two core measurements — depth-weighted playing time and a depth index at control points — adjusted by structural penalties.

The same methodology is applied to every structure without exception. No tournament receives a different scoring model because of its brand, location, organizer or buy-in.

Exact formulas, weightings and thresholds remain proprietary in order to protect the integrity of the rating system.

Estimating the average stack

The number of entries is not known at the time of rating, and it is not needed. The average stack is derived from the starting stack and the fraction of the field still in play; the number of entries cancels out of the calculation.

The remaining field fraction is estimated by a parametric survival model, calibrated on actual player counts from major international tournaments recorded after registration closed. The model is identical for every structure: the difference between two scores reflects only the difference between two structures.

The model is indexed on a blind-pressure clock rather than on the level number: the clock advances with the relative increase of the big blind. A repeated level does not move it, and the effect of a sharp blind jump on the clock is capped.

The average stack is measured at the start of each level, when the new blind takes effect.

The model’s pace does not depend on how levels are numbered.

Score components

Depth-weighted playing time — primary component. Each level is assigned to a depth stage according to the average stack expressed in big blinds, from free play to critical play. Its duration is weighted by a stage-specific coefficient: full weight in deep play, decreasing thereafter, zero in the critical zone. The total is aggregated through a diminishing-returns function.

Depth index at control points — secondary component. The average stack is sampled at several control points distributed along the clock, with maximum weight in the middle of the tournament. Each reading accounts for the real cost of an orbit — small blind, big blind and ante — and for the number of players per table, in the manner of effective M: it measures how many hands the average stack can fund. Each reading is first weighted by the length of the level — the longer the level, the more the depth is worth, with no ceiling — then converted into points on a logarithmic scale.

Penalties

Penalties are applied multiplicatively and are capped.

Inflection points: any blind increase above the tolerance threshold, in the central phase of the tournament, triggers a cumulative penalty.

Late registration close: the penalty depends on the residual depth of a player entering at the last registration level.

What GSI does not measure

GSI does not measure the prestige of the event, the prize pool, the number of entrants, the organizer, the casino, the field or the player experience.

GSI measures the tournament structure — not the event around it.

Does the buy-in affect the GSI Score?

No.

A €100 tournament and a €10,000 tournament are measured using exactly the same structural rules.

The buy-in amount has zero weight in the GSI rating.

The GSI Score

Every complete and validated structure receives a GSI Score from 0 to 100.

The score is accompanied by a seven-star visual scale with proportional fill and a label ranging from Turbo to Championship.

The detailed GSI report explains the structure’s measured strengths and weaknesses and provides a level-by-level analysis.

The numerical GSI Score remains the primary rating.

Calibration and scale

The scale is calibrated on a panel of reference structures: recognized international and French Main Events. Reference values are frozen and any change is traced; they do not change from one import to the next.

Beyond the reference, the gain curve is asymptotic: the maximum score is never reached, which preserves headroom for any structure that outperforms the panel. At the low end, the output keeps resolution on fast structures, without compressing them toward zero.

The final score is obtained through a strictly monotonic transformation of the points: the relative ranking of structures is preserved.

One frozen engine. A reproducible score.

The GSI Intelligence Engine is frozen for the season. Within that season, the same structure must produce the same score.

A future season may use an updated engine. If the methodology changes, the new engine is applied consistently to every structure in that new season.

Closed seasons are never recalculated with a later engine. A 2026 score remains a score produced under the 2026 methodology.

A published score may also change if the underlying structure data is corrected or completed.

A standardized scope: one common evaluation window

Every structure is measured over the same evaluation window, defined on the blind-pressure clock, excluding breaks. Later levels are not included in the score.

This common window gives a standardized basis for measuring structures of different formats and durations under the same rules. It neutralizes the effect of an artificially lengthened schedule and avoids giving disproportionate weight to very late levels that only concern a small part of the field.

Rebuys, re-entries and add-ons are not taken into account: only the starting stack is used, so that the structure is scored independently of the entry policy. In a re-entry tournament, the chip mass per remaining player is higher than in a freezeout; this effect is deliberately excluded.

GSI states this scope clearly: the score measures the defined evaluation window, not the entire published blind schedule.

When a structure is not scored

A structure is not scored when its schedule is too short to be measured, or when its game format falls outside the scope scored by GSI.

A schedule that stops while residual depth remains significant receives the status “incomplete schedule”.

Missing levels are never extrapolated.

Equal measurement means equal score

If two structures are identical over the evaluation window, everything GSI measures is identical. Their GSI Score is therefore identical, even if their blind schedules differ later.

This is deliberate. The same measured blind structure — whether used in a satellite, qualifier or recurring tournament — must produce the same rating.

A rating that changed between two identical measured structures would not be reproducible.

How ties are ranked

Structures with the same GSI Score share the same rank, following standard competition ranking.

If two structures are tied for first place, both rank first. Second place is skipped and the next structure ranks third.

Breaking a genuine tie would mean inventing a difference GSI did not measure.

Same measurement. Same rating. Same rank.

Validation

The GSI Intelligence Engine has been validated by a battery of consistency tests across every structure in the database: chip-scale invariance; monotonicity with respect to starting stack, level duration and late registration close; insensitivity to levels added outside the window.

Known limitations

The average stack is a central statistic, not the stack of any given player; since the convention is common to every structure, it does not bias the comparison.

The survival model describes a typical tournament. The actual elimination speed of a given tournament may differ, with an effect of the same order on every structure.

What the GSI Method gives players

GSI turns a published poker tournament structure into a score that can be understood within the same methodology used for every structure.

That makes it possible to explore rankings and understand how each structure is measured on the same measurement basis — without changing the scoring rules from one tournament to another.

Demonstration example

Illustration only. This tournament does not exist and is included in no GSI index, ranking or award.