Mana Games

[nickkyrgiosfan]

Hello. For a while I have been trying to figure out why I feel that TE4 is still fast paced when compared to real life, and I think the origin of the issue is that the court friction is very fast. I have spent a long time measuring the exact speed of balls before and after bounces in this game and I see that even when you replace any court's CoF value with a really high number, the court speed is still faster than most tennis courts IRL. I've replicated lots of serves and groundshots from real life clay courts like the French Open into TE4 and there is a very noticeable difference, even when I set the COF to its max value which should make it as slow as possible

I remember in TE 2013, the court.ini allowed you to edit the friction value to whatever you wanted, it allowed you yo make the court really fast or really slow but it also allowed you to perfectly replicate the speed of real clay courts through a low value. In TE4, you can make the court very fast, but not very slow like you could in 2013. You can't make the court speed slow enough to match real clay courts because 0.99 is the maximum.

[manutoo]

Hello,

how did you make your measures ? What exact results did you get ?

What clay surface did you use in-game ? (What is its exact name)

[nickkyrgiosfan]

The surface I am using right now is RG Philippe Chatrier Day, from the XKT mod, but everything I'm saying is general and applies to the vanilla clay court as well

The CoF value I am using is 0.999 (anything above that does not seem to do anything but strangely affect the players' behavior)

I measured using a software called Kinovea to allow me to measure the time frame by frame from a video. What I did is, I found videos from RG aces and winner grounshots, then replicated them in the game with the exact speed. I used the software to confirm the speed matched, because the time it takes for the ball to hit the ground after leaving the racket, matches the RG videos. However, the time it takes for the ball to bounce off the ground and hit the court again, does not match. In the game, the shot lands faster than in real life. You can also see at first glance that the post-bounce speed is noticeably faster in the game. This shows to me that even though the shots and speed are perfectly replicated, the court does not slow down the ball enough to match the real RG courts. (Even at the highest allowed CoF)

Is there any possibility of you allowing us more freedom in modding the CoF to try and experiment to make the courts slower, kinda like in TE 2013 ? It would be very useful

[manutoo]

Thanks, but this doesn't actually measure the court friction.

The time between the 1st and 2nd bounce mostly depends on the vertical velocity after the bounce, ie: how high the ball bounces. The horizontal post-bounce speed mainly changes the distance between the 2 bounces, not the time between them. So if the 2nd bounce happens sooner in TE4, it doesn't show that the court is too fast; it may simply mean that the ball bounced lower. Likewise, matching the time between the racket strike and the 1st bounce doesn't guarantee that the trajectories are identical. The launch angle, strike height and spin also matter.

To check the CoF, we'd need to compare the horizontal ball speed immediately before and immediately after the bounce, ideally with the same incoming speed, angle and spin.

Could you measure that in Kinovea for one or two of your examples ? That would give us something meaningful to compare.


Also, the XKT "RG Philippe Chatrier Day" surface and the vanilla clay aren't actually identical:

XKT RG Philippe Chatrier Day:
CoF = 0.75
CoR0 = 0.84
CoR1 = 0.78
DeltaE = 0.01

Vanilla clay:
CoF = 0.76
CoR0 = 0.85
CoR1 = 0.78
DeltaE = 0.02

So there are already small differences not only in friction, but also in vertical restitution and topspin elasticity.

If your method doesn't reveal any meaningful difference between these two surfaces, then it's probably not precise enough yet to isolate the effect you're trying to measure.


Also, unlike TE2013, which used a simplified friction model, TE4 uses a physically based Coefficient of Friction (CoF), with values based on official ITF measurements of real tennis courts. These measurements already account for the actual loss of horizontal ball speed during the bounce, including the effects of the court surface.

So I have good reasons to believe the current physics are reasonably accurate. Of course, that doesn't mean they're necessarily perfect, but I'd need much more precise measurements, comparing the actual ball velocities before and after the bounce under similar conditions, before concluding that something is wrong.

And simply allowing a CoF above 0.999 wouldn't necessarily help, as increasing friction beyond a certain point can have little or no additional effect on the bounce physics. 🙂