The Sony is as good as it gets, and without confirming what was “actual” (through video analysis for example), versus what each sensor said would be the only way to check accuracy. Otherwise you’re, as you said, “guessing”.

For our test – the devices both tracked around 600 shots – we provided a % breakdown of shots played to give an indication of the metrics. I am guessing – during a match – there are lots of shots which are not really in-game shots – you may for example just lob the ball over the net to give your opponent a tennis ball to serve with. Babolat may be better at weeding these out. I think the reason Qlipp showed a higher % of forehands – its because it sometimes misinterprets a serve as a forehand. Also – this is why the Babolat % does not add up to 100% – we excluded the serves (the Qlipp didn’t recognise very many).

In my view – this is the first generation of these devices (apart from Babolat). Its far from being an exact science at this stage – and we have to be realistic in our expectations. Maybe in 2-3 years, when they start monitoring your movement in 3d you can compare shot by shot how good they are. For now, this is as good as it gets.

Sorry but those comparison numbers look wildly different to me. The Qlipp registered 78 more shots than the Babolat Play is what that tells me?! Also, you’ve shown what shots were registered by the devices, but not whether that was in fact the shot played. Unless I’m misreading something.

Hi – we did. The table where we compare it to a Babolat Play Connected can be seen above. You can see the breakdown is fairly similar. Qlipp does struggle though to recognise serves.