About ATTR

Why this exists

Tennis is one of the few sports where cross-era comparisons are a daily obsession — yet there is no standardized tool to make them honestly. Is Djokovic better than Federer? Was Sampras the best of the 90s? How does Alcaraz at 21 compare to Nadal at 21? These questions deserve a rigorous answer, not an opinion dressed as a fact.

The problem is structural. The ATP point system was never designed for historical comparison — it was built to decide who plays in which tournament next week. It resets every 52 weeks, has changed its scoring rules multiple times since 1968, and treats a title in a depleted draw identically to one against the full strength of the tour. Without a fixed reference system, every cross-era comparison is noise.

ATTR was built by a data engineer with a simple conviction: the sport deserves better than this. A single, consistent score that adjusts for context, standardizes across rule changes, and lets you compare careers across generations — not just seasons.

The Score

Every result since 1990 is assigned a career score point. The ceiling is set by tournament prestige — a Grand Slam win is worth more than a 250 win, always. But the ceiling is then adjusted downward by two independent multipliers that capture something the ATP never measures: how hard was it really?

All three factors sit between 0 and 1 relative to the prestige ceiling — the final score is always ≤ the raw prestige value. A 3-0 win over the world #1 in a stacked draw approaches the ceiling. A walkover in a depleted bracket does not.

Prestige Points

Fixed weights per tournament tier and round — never adjusted for inflation, never repriced between seasons. A Grand Slam title in 1993 and one in 2023 both start at 2,000. This is the one constant in the system.

Example: Sinner wins the Australian Open 2026. Prestige = 2,000. This is the maximum possible — the two multipliers below will bring it down based on draw quality and match dominance.

SPS — Seed Presence Score

The SPS captures draw quality — not as a binary "was the field good or bad" judgment, but as a continuous score derived from the actual ranking of every opponent beaten along the way. A logarithmic formula maps each opponent's rank to a value between a floor and 1.00, then averages across all rounds played.

The floor and max_rank increase as the tournament progresses — surviving deeper into a Slam is harder, so the penalty for beating a lower-ranked opponent is smaller in the final than in the round of 16.

Popyrin wins Montreal 2024. He beats opponents ranked 36, 12, 4, 1 across the tournament. His SPS reflects the actual quality of each round — not a generic field assessment. Final SPS ≈ 0.82.

Pre-2000 data lacks opponent ranking granularity. Those tournaments default to a neutral SPS of 0.80 — conservative, not penalizing.

Set Multiplier

How you win matters. Dominating 3-0 is not the same as grinding 3-2. The Set Multiplier captures match dominance for the winner, and resilience for the loser — because reaching a fifth set against a top player has value too.

Sinner loses the US Open 2025 final 1-3. Prestige = 1,200. SPS = 0.6797. Set Multiplier = 0.94 (loses 1-3). Career Score = 1,200 × 0.6797 × 0.94 = 766.

Scope & limitations

• ·Coverage starts in 1990 — better data, structural equivalence to the modern tour
• ·Team events (Davis Cup, Laver Cup, United Cup) are excluded
• ·Pre-2000 SPS defaults to 0.80 — opponent ranking data not granular enough
• ·Walkovers and retirements counted at face value — a known limitation
• ·Data updates weekly. ATTR is independent — not affiliated with ATP or ITF