Personal tennis stats

O

olivierd

New User
After a tennis match with a friend of mine, he mentioned that he'd love to be able to get statistics about the match he plays (i.e.: Aces, Double faults, 1st serves in/out, 1st serve points won, 2nd serve points won, Net points won, Receiving points won, Winner shots, Unforced errors, Forced errors, Return winners, Return unforced errors, Return points, Return points won, forced errors, unforced errors by play type and hand).

I'm curious if other players feel the same way. I'm looking for a side project to work on and wondered if that is something a lot of players are looking for. And whether players would be willing to pay for such a service.

What do you think?
 
You have to use something like PlaySight. Even the sensors on the racket don't give you this kind of information but just stroke information. Just video recording won't do it either. You need Playsight combined with a superb cognition software to analyze the match, and building this software will be very difficult.
 
I use an app called Tennis Tracker Pro to track junior matches. You enter the info either watching it live or on video. It has a few shortcomings, but by inputting certain data you get:

Match duration
Total points won
1st serve percentage
Aces
Double faults
Break points saved
Points won on own 1st serve
Breaks points converted
Points won on opponent 2nd serve
Winners
Unforced Errors
Forced Errors
Forehand Stats
Backhand Stats

You can view as a list of stats as well as charts and timelines within the match. Really useful to find patterns and weaknesses to work on in practice. Only wish it had service winners/return winners and net stats.
 
Thanks for the response guys. I did not know about playSight. it seem to be doing everything I wanted to do for this project except for its high price tag. I'd want to figure out a way to do it so that it is accessible to a wider range of players...

BMC9670, how much of a pain is it to input all this data. I'm assuming someone does it for you during your games. Or do you do it yourself?
 
olivierd said:
Thanks for the response guys. I did not know about playSight. it seem to be doing everything I wanted to do for this project except for its high price tag. I'd want to figure out a way to do it so that it is accessible to a wider range of players...

BMC9670, how much of a pain is it to input all this data. I'm assuming someone does it for you during your games. Or do you do it yourself?
Click to expand...

I wouldn't call use of humans a solution in this case. It is just unfeasible other than for juniors or pros whose coaches or parents do it for them. I suppose an app which permits fast entry will be useful.

But I am surprised that you thought that you could get the same functionality as Playsight in your project in the first place. It is very advanced and illustrates the point that what is easy for humans (identifying an ace, e.g.) is very difficult for a computer (and vice versa). That is the whole challenge of Artificial Intelligence.
 
Thanks for the input. The point of my project is to get something similar to play sight but in a more flexible and cheaper solution. I'm hoping to come back to this thread in a few month with a link to a product;-)
 
olivierd said:
Thanks for the response guys. I did not know about playSight. it seem to be doing everything I wanted to do for this project except for its high price tag. I'd want to figure out a way to do it so that it is accessible to a wider range of players...

BMC9670, how much of a pain is it to input all this data. I'm assuming someone does it for you during your games. Or do you do it yourself?
Click to expand...

Unless it's tracking technology like Playsight - which for the time being is expensive and involves precise set-up on court, data has to be entered by a person. With a tracker app, you have to enter it. I do it as I'm watching matches as it gives me something constructive to do while watching. I've also entered data from video of matches. It's very easy and fast to enter the data.

I guess I misunderstood what you are trying to do. I thought you wanted to provide a stat service where people would send you video of matches and you send them stats for a fee. I guess if that's the case, they can buy a $10 app and do it themselves much cheaper.
 
sureshs said:
I wouldn't call use of humans a solution in this case. It is just unfeasible other than for juniors or pros whose coaches or parents do it for them. I suppose an app which permits fast entry will be useful.
Click to expand...

See my post above - I thought he was thinking about a data entry service based on video of matches. Coming up with automatic technology like Playsight is another ballgame altogether.
 
olivierd said:
Thanks for the input. The point of my project is to get something similar to play sight but in a more flexible and cheaper solution. I'm hoping to come back to this thread in a few month with a link to a product;-)
Click to expand...

Playsight has both hardware and software. Without the high-speed cameras, there would be no data to process. Now, someone could invent such powerful algorithms that a single camera could be enough while allowing for some percentage of wrong decisions.
 
BMC9670 said:
See my post above - I thought he was thinking about a data entry service based on video of matches. Coming up with automatic technology like Playsight is another ballgame altogether.
Click to expand...

That is what I first thought, but that seemed too trivial to me (data entry app). Moreover, how can it be a solution to his problem about his match with a friend (which implies recreational play)? Which club player will bring a tennis caddie with him to take notes?

And it did turn out that my guess was correct and he wanted more - "playSight. it seem to be doing everything I wanted to do for this project "
 
Playsight has both hardware and software. Without the high-speed cameras, there would be no data to process. Now, someone could invent such powerful algorithms that a single camera could be enough while allowing for some percentage of wrong decisions.
Click to expand...

Gonna be hard to beat playsight with a single camera solution that not permanent....
 
This is possible, even using a single camera. I think the main difficulty is we would have to handle a variety of visual interference and widely varying conditions where we recreational players play games. Are you trying to do this in real-time? If so, the level of effort is significant.
 
A single camera cannot theoretically overcome the obstacle problem. Some "intelligent guess" mechanism has to be added which will be wrong in some cases.

Hawk Eye is used for making line calls so it has to be more accurate and multiple cameras are a must. If line-calling is not needed, simpler hardware mat be possible, like a single camera.
 
The idea would be to use 1 or 2 camera to film the game and to provide the stats after the game.
For now, i'm not targeting line calling or real time stat generation because of the hardware which would be required as other have pointed out (although one could argue that in a couple years, that hardware will be widely available :)).

The idea here is not to beat playsight. I think playsight has its market very well covered, but that market is limited because of cost (which itself will come down with time too). As other have pointed out, i'd like to create something more accessible.

The real question is: would some of you be willing to pay $10 to get one of your tennis game annotated and get your stats by simply filming your game with say your phone or a goPro?

BMC9670 has already answered in the negative ("I guess if that's the case, they can buy a $10 app and do it themselves much cheaper."). But what about the others?

Thanks for all the input!
 
olivierd said:
The real question is: would some of you be willing to pay $10 to get one of your tennis game annotated and get your stats by simply filming your game with say your phone or a goPro?
Click to expand...

This is what I thought you were meaning to do. I guess it would be worth paying $10 to save the time of compiling the stats themselves from the video. On the other hand, if the fee is only $10, you certainly won't be making enough money to make offering the service worth your time if a typical match is, say, 1.5 to 3 hours long. Tough model to make work, IMO.
 
The software development is formidable. You need object recognition, scene recognition, virtual 3D world with annotations, motion detection, image processing, video processing and so on. These have taken decades to develop. Some of it is available in open source libraries put out by Universities, but they will typically be not legally available for commercial use without paying royalties.

Just putting together the software pieces after that will be a lot of work.

And the final product will be so computationally intensive that it may not run on a phone, without extensive optimization based on use of GPU and image engines through OpenCV API.

Check this out:

https://www.qualcomm.com/products/vuforia
 
Its already been mentioned but use play sight.....its the best thing out there with little effort. Find a club that has it you can search their website
 
I would pay $10 per match to have my match stats recorded =) I would be okay if it were not real-time also, e.g. upload my video for processing in the cloud, after the match.

Also for reference you can take a look at Google Project Tango: https://www.google.com/atap/project-tango/about-project-tango/
If depth perception becomes commonplace for personal mobile devices in the near future, you may be able to leverage that. I'm not sure about the limitations of the depth sensing though, e.g. range, capture rate, etc.
 
There's no such thing as "statistics". It's a storytelling device, like in football. It part of broadcasting and storytelling to recall statistics and tell possible matchups, tendencies, style make fights, so on. Statistics is invented for storytelling.
 
counterfeit25 said:
I would pay $10 per match to have my match stats recorded =) I would be okay if it were not real-time also, e.g. upload my video for processing in the cloud, after the match.
Click to expand...


I would pay someone $20 to chart a full match... as long as they knew what they were doing, and we're able to chart just about everything the OP listed initially.

I wouldn't do it often... I think one or two matches vs different opponents would give you a very good picture of what your strengths/weaknesses are... not that we don't already have an idea, but it could show some blind spots we have towards our own game, or affirm things we do well.

Kinda like the first time you see yourself play on video. Very enlightening and revealing.
 
Congrats on the website and project. A buddy of mine and I would keep stats for each other during matches and it's helpful. Not sure I would pay a monthly fee for this service but it's good you offer an "occasional" option for users. How long does it take for a user to get their stats and analysis sent back to them?
 
You must log in or register to reply here.

Similar threads

Kobble
Strange match stats. If this was a real world tennis problem, I think a player might melt down on court.
Replies
4
Views
822
Fairhit
Fairhit
Moses Man
Some match play and hitting to critique!
Replies
4
Views
1K
Jackie T. Stephens
Jackie T. Stephens
K
Stats for 1990 Wimbledon final (Edberg-Becker)
Replies
16
Views
4K
abmk
abmk
M
Stats for Sampras-Fed Macau 2007 Exhibition
Replies
3
Views
943
helloworld
H
W
Match Stats/Report - Lendl vs Becker, Queens 1990
Replies
9
Views
2K
TimeToPlaySets
TimeToPlaySets
Back
Top