Setpoint Data Labsby GFI AI Software Contact us
Setpoint Data Labs by GFI AI Software, Hong Kong

We compute
tennis.

Setpoint is the tennis division of GFI AI Software. For years we have been collecting, cleaning and merging tennis data into one database. Now we build products on it: rankings, news, an encyclopedia, a data API and video analysis for coaches.

19.2M matches since 1877 1.7M players 5 products on one database 12 languages
/ 01 · What we do

What we have
built over the years.

Everything starts with the database. We gather results from the tours, federations and junior circuits, check them against each other, remove duplicates and link every match to the right player and tournament. Then we add what nobody else has: our own texts, portraits, live scoring and a strength rating.

01

One consolidated database

19 million matches since 1877, 1.7 million players and close to a million tournament draws. ATP, WTA, Challengers, ITF, juniors, college tennis and children's tournaments in one structure, cleaned, cross-checked and labelled by us.

02

Our own content

Biographies of players, descriptions of tournaments and of the events that shaped the sport, written by us and translated into 12 languages. Player portraits drawn by us, so they can be used in other products without rights problems.

03

Live scoring, point by point

We do not resell live feeds. We follow each match ourselves and keep the score point by point, for almost every match in play. When equipment on a court fails and points go missing, our method reconstructs them mathematically, so the record stays about 99% accurate.

04

TSR, our strength rating

Tennis Strength Rating is our own measure of how strong a player is right now, computed from the full database and refined with professional players and coaches. It shows how well a ranking position reflects real strength.

05

Match video analysis

A model that reads a match from a video file or a stream: strokes, errors, positions, patterns. It turns what it sees into a report, a training plan and advice for the next opponent.

/ 02 · Technology

How the
data moves.

→ 01
Collection

Results, draws, rankings and live scores from the tours, the ITF, national federations and junior circuits, plus our own scoring of matches in play.

→ 02
Cleaning and merging

Every record is checked against the other sources, duplicates are merged, names are matched to one player card and tournaments to one calendar. This is where most of the years went.

→ 03
Computation

Live scores, game and match situations, the TSR rating and win probabilities are all computed from the same database, after every finished match.

→ 04
Delivery

The same computed data goes to rank.tennis, today.tennis, wiki.tennis and analytics.tennis, and to customers through api.tennis.

INGEST 400+ sources NORMALIZE 1 schema COMPUTE point-by-point DELIVER <800 ms PRODUCTS × 5 live court
Stack Python · ClickHouse · Kubernetes · WebSocket · ONNX
/ 03 · TSR

Tennis Strength Rating
shows real strength.

An official ranking counts the points a player earned over 52 weeks. TSR answers a different question: how strong is this player today, and how likely is a win in the next match.

TSR is our own system, and we know of no equivalent. It is computed from the whole database, 19 million matches from Grand Slam finals down to children's tournaments, and it was tuned together with professional players, coaches and analysts who told us what the numbers were missing.

One scale from 0 to 100 covers ATP, WTA, ITF, juniors and children's tennis, so players from different circuits can be compared directly. The rating is recalculated after every finished match, and with it the win probability for every upcoming match.

The production model is TSR 6.5. In public we show a lighter version, TSR Lite, on wiki.tennis, rank.tennis and through api.tennis. It is enough to see where a ranking position overstates or understates a player, and to get a probability for any match before it starts.

0·100
One scale for every circuit, from ATP to children's tennis
19M
Matches the model is built on
~75%
Singles matches predicted correctly in the back test
6.5
Current production model. TSR Lite is the public version
/ 04 · Products

Five products
on one database.

Each product shows a different side of the same data. Some are free, some are commercial, and a correction in the database reaches all of them.

/ 05 · Founded from inside the tour

A rare synergy
of IT and sport.

Setpoint Data Labs was started by two people who came to tennis from different sides. One is a veteran IT executive who has spent 25 years building and turning around technology companies. The other is a professional tennis player, WTA doubles world No. 14, who spent her career on the tour.

She had seen for herself how poor the data in professional tennis was: feeds arrived late, points went missing and the stats often didn't hold up when you checked them. He knew how to build the system that fixes that. Setpoint is what came out of it.

Today Setpoint runs five products on that database, free and commercial, and keeps adding to it every day: new matches, new texts, new portraits.

Setpoint Data Labs is the tennis division of GFI AI Software Limited (Hong Kong). GFI AI Software owns the database and the today.tennis, rank.tennis, wiki.tennis, api.tennis and analytics.tennis services.

“Everything you see in our products was computed by a system we built, from events we collected ourselves.”
Founders of Setpoint Data Labs by GFI AI Software
Irina Khromacheva
Irina Khromacheva
Co-founder · Product
  • · WTA doubles world No. 14
  • · WTA 1000 champion (Wuhan 2024)
  • · 12 WTA doubles titles
  • · Former junior world No. 1
Maxim Khromachev
Maxim Khromachev
Co-founder · Strategy
  • · 25 years in IT, 10+ launched projects
  • · CEO-level IT management, security, large teams
  • · Business experience in 167 countries
Products live: 5 Sales start: December 2026 Live assistant for coaches: in pilot The database grows every day
0M
Matches in the database, since 1877
0
Players with full profiles
0+
Tournaments & draws
0M
Point-by-point rallies
~0
Junior players profiled: Europe, Asia, LatAm, Australia, Japan

We're hiring data engineers, ML engineers and tennis analysts.

We are a small team building real-time systems for a sport we love.

View open roles →
/ 06 · Contact

Build with
our data.

[email protected]
Setpoint Data Labs by GFI AI Software, Hong Kong