Hacker Newsnew | past | comments | ask | show | jobs | submit | RossTennis's commentslogin

Would you have an email address or another way of contacting you? Your contact form / suggest a game links appear inactive at present. Would love to discuss something with you.


Hi folks,

After a couple years of toil and grind I finally managed to build a predictive model for pro tennis. It analyses around 100 data points per player and creates odds / probability of winning the match. That model is now winning at a flat rate of 12.5% (over a smallish sample of a few hundred bets, but looks solid).

So, after the model was working, I decided to build a website with my friend where other tennis fans can check matchups, see predictions and get a better head to head analysis than anything else that exists out there. (Point: Most other h2h sites show data which is not directly comparable between the two players, where as my site produces relative comparison points).

I'd love for you to check out the site and see what you think. Any feedback, good or bad, is always welcomed! And if you are a tennis fan you can join my whatsapp betting group too.

And if you are interested in the business side of things:

We are trying to generate revenue three ways: 1) affiliate revenue 2) licensing our data to other sites 3) the betting fund (currently 25k)

We're also working heavily on SEO optimisation, and building up a community of tennis betting fans.

I might also sell tips in the future as there is a big market for that but I'm currently aiming more high end, e.g. fund investment, high priced tips. Pretty much testing a few things at present and see what might work.

Thanks for checking it out.

Ross.


I spent the last year building a predictive tennis model using Google Sheets which is basically the same as Excel. I have no technical skills so this was the easiest way to do it, plus it helped me learn SQL. Basically I have around 50,000 tennis matches in one sheet, filter them into another sheet, into around 100 comparative points, then in the end produce a probability for each upcoming match, and bet accordingly. I also use it to produce a little head to head stats visual and automated 200-word match previews.

It all works perfectly but obviously just for one user at a time. So I brought on a dev friend a few months ago and will launch a user facing site next week, as well as a b2b api for match previews / seo rich content.

I'm actually looking for seed funding up to around 25k, if anyone wants to reach out about that, or has other questions.

If you just want to follow some of the tips I'm running a very basic blog site at www.tennisacca.com sharing my most highly recommended picks (rather than listing 100's of games a week and needing a huge bankroll), which win at a ROI of 10%+ on average. It's mostly just for friends until the main site launches next week.


Hi! I did something similar a few years ago,instead I used python and machine learning algorithms + custom ELO implementations for tennis matches. Unfortunately it worked too well and I was banned from all the bookmakers.


nice, well done! Why not use betfair or Molly? Or also sell tips etc? Do you have a site?


Out of curiosity. How did you manage to gather so much data?


Various data feeds, working in the industry, plus some open source


I've actually been building a tennis model to beat the bookies for around a year or so and am now beating them by ~6% overall. About to launch a public site in next few weeks. In the mean time, if you'd like to follow picks for free for the US Open just drop me a line and I'll add you to my group.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: