A social network for sports predictors
- Timeline
- MVP — 4 months
- in release in under a year
Task: a public platform where an author publishes a prediction with their reasoning, and the result is visible to everyone.
How it works
- 01Prediction feedThe match, the market and the odds, the author’s reasoning, and repeat and share
- 02Author pageBank, profit, ROI and win rate, the settled predictions and the subscribers
- 03Top predictorsA ranking by ROI, profit, average odds and win rate over 30 and 90 days
- 04Main storefrontPredictions from authors worldwide, filtered by sport and by community
20 000 active users a day · 400 rps at peak · 10 unique screens
Particulars and difficulties
A claim has to be settled
An author’s figures are computed from settled predictions only — that is what makes a subscription mean anything
A rank cannot be bought by volume
The ranking is built on return rather than on the number of posts
A peak load on a public feed
20 000 users a day and 400 requests a second at peak
Ten screens on one component
The prediction card is the piece the whole product is assembled from
Results
- 20 000
- active users a day
- 400 rps
- at peak
- 10
- unique screens
Following means something
The result of a prediction is visible to everyone afterwards, so an author is judged on the distance
In release in under a year
MVP in four months and the full platform inside the same year
Stack
PHP · Symfony · PostgreSQL · Redis · Elasticsearch · Google Gemini