Forehand Timing: Your Contact Point Is Decided Before the Ball Bounces

Here is the claim, stated plainly: you cannot fix a late forehand by reaching further out in front. That instruction is the most commonly offered advice about forehand timing on any public court, it is aimed at the right symptom, and it does almost nothing about the cause. Contact point is not a place you steer the racquet toward. It is the output of a sequence that began roughly a second earlier, most of it before the ball bounced, and by the time the racquet is anywhere near the ball, the decisions that mattered have already been made.

The rest of this piece is spent earning that sentence. We'll do it with the numbers, because the numbers are what make the problem legible — and once you see how small the target actually is, the frustration of hitting well on Tuesday and badly on Thursday starts to look less like a character defect and more like an arithmetic one.

The target is smaller than almost anyone assumes

Start with the impact itself. Rod Cross, the physicist at the University of Sydney who has spent decades measuring what tennis balls and string beds actually do to each other, puts ball-on-strings dwell time at roughly 4 to 6 milliseconds for a typical groundstroke. That is the entire duration of the event we call "contact." Nothing you decide during it matters. Nothing you can perceive during it matters either — the visual-motor loop is one to two orders of magnitude slower than that.

So the real question is not the instant of contact but the window that produces it. Suppose a rally ball at the 3.0–4.0 level is traveling around 18 meters per second after the bounce — call it 40 mph, which is unglamorous but realistic for a ball that has already lost energy to the air and the court. Suppose further that the depth tolerance of a good contact zone is about ten centimeters either side of ideal: closer than that and the elbow jams, further and the arm runs out of extension while the face is still opening.

Twenty centimeters of tolerance, divided by eighteen meters per second, is about eleven milliseconds.

That is the width of the window. Not the width of the racquet, not the size of the sweet spot — the amount of time the ball spends inside the volume where the stroke works. Eleven milliseconds is roughly a tenth of a blink.

The consequence runs the other way too. If the whole sequence starts late by some amount, the ball keeps approaching while the racquet is still on its way, and the two meet somewhere closer to the body. The shift isn't the full ball speed times the delay, because the racquet is also traveling — the meeting point moves at a rate set by both.¹ With an 18 m/s ball and a racquet head at 20 m/s, the contact point retreats about 9.5 centimeters for every 10 milliseconds of lateness.

The assumptions are stated above and the geometry is simplified to a head-on approach; a crosscourt ball arriving at an angle changes the numbers somewhat but not the order of magnitude. The point stands: a fiftieth of a second is the difference between clean and cramped. No conscious correction operates at that resolution. Which is why the correction has to happen upstream, where the units of time are tenths of seconds instead of hundredths.

Where the milliseconds actually get spent

The sequence, in the order it happens

The opponent's swing. Information about where the ball is going exists before their strings touch it. This is one of the better-replicated findings in sport perception: temporal occlusion studies, in which video of an opponent is cut off at various points and viewers are asked to predict ball direction, consistently show that skilled players extract usable directional information from pre-contact body kinematics. Damian Farrow and Bruce Abernethy ran a series of these in tennis through the early 2000s, and the broader anticipation-training work by Mark Williams, Paul Ward, Nicholas Knowles and Nicholas Smeeton, published in the Journal of Experimental Psychology: Applied in 2002, showed that this skill improves with structured video training and transfers, at least partially, to on-court performance. Their training groups were small, as these studies almost always are.

Your split step. It should land as the opponent strikes, not after. A split that lands 100 milliseconds late doesn't cost you 100 milliseconds at contact — it costs you that, plus every downstream step it delays.

The turn. Trunk and shoulders rotate as one piece, racquet coming with them. Reid, Elliott and Crespo's 2013 review of forehand mechanics and learning practices in the Journal of Sports Science and Medicine describes the stroke as a proximal-to-distal sequence: large segments accelerate first, smaller ones later, each borrowing momentum from the one before. The turn is the cheapest timing insurance available, because it is the only action in the entire chain that does not require knowing exactly where the ball will land. Direction alone is enough to justify it.

The bounce. This is the anchor, and there is a striking piece of evidence for why. Land and McLeod, writing in Nature Neuroscience in 2000, tracked the eye movements of three cricket batsmen and found that they did not smoothly follow the ball. They made a predictive saccade to where the ball was about to bounce, waited for it there, then tracked briefly afterward. The better the batsman, the earlier the saccade — by about 100 milliseconds. Three subjects is a very small study, and cricket is not tennis. But the structural logic transfers: the bounce is where the ball's remaining trajectory becomes knowable, and the visual system appears to organize itself around getting there first.

The last 200 to 300 milliseconds. From bounce to contact on a typical rally ball, that's all there is. And this is where the most interesting result in the literature lives. Bootsma and van Wieringen, in the Journal of Experimental Psychology: Human Perception and Performance in 1990, filmed five expert table tennis players hitting attacking forehand drives. The variability of their timing at the moment of contact was far smaller than the variability of the movements that produced it — on the order of a few milliseconds at contact, against much larger fluctuations earlier in the stroke. The movement was not being replayed from memory. It was being steered, continuously, right up to the end.

Five players, one sport, thirty-five years ago. But it reframes the whole problem. The stroke is a control process with a deadline, not a recording. And control processes need runway.

Where should you make contact on a forehand?

Ahead of the trunk, at about waist-to-chest height for a standard drive, far enough from the body that the elbow stays comfortably bent rather than either jammed or locked. That's the honest short answer, and the reason it isn't a number in centimeters is that the published numbers move around considerably depending on stance, shot type, and player height. Bahamonde and Knudson's comparisons of open- and square-stance forehands in the early 2000s found meaningfully different joint kinetics between the two, which means "correct contact point" is partly a function of which stance you're in.

The invariant worth holding onto is relational rather than metric: contact happens ahead of the plane the chest is facing, while the torso is still rotating forward. That phrasing matters, because it makes the reference frame explicit. If the torso is late, the arm can be as far in front as it likes and contact will still be behind the body's forward face. "Out in front" is meaningless without specifying in front of what.

Why reaching forward makes it worse

Now the claim from the opening can be cashed out.

When a player who is arriving late decides to hit further out in front, the only segment still available to them is the arm. The torso has already committed. So the correction becomes an early extension of the elbow, and three things happen at once.

The arm reaches its extension limit sooner, so racquet head speed at contact drops — the distal segment is no longer being whipped by a rotating proximal one, it's being pushed. The chain that Reid, Elliott and Crespo describe runs in the wrong order, and the elite-versus-sub-elite kinematic comparisons in the literature, including Landlinger and colleagues' 2010 study in Sports Biomechanics of crosscourt and down-the-line forehands, consistently locate the difference between skill levels in exactly this sequencing rather than in effort.

The second problem is geometric. The racquet travels an arc, and the face angle and velocity vector both rotate along it. Moving contact forward doesn't just move it forward — it moves it to a different point on that arc, where the face is oriented differently. A player who reaches is not making the same contact earlier. They are making a different contact.

The third is that it is a conscious correction applied to an eleven-millisecond window. Even granting Bootsma's finding that late corrections exist, the correction machinery is perceptual and continuous, not deliberate and verbal. You cannot issue an instruction to it.

So the advice is aimed at the right target and delivered to the wrong address. Contact out in front is a description of good timing, not a method for it.

The two mirror-image errors

Most timing failures we see described in coaching literature fall into one of two shapes, and they are opposites, which is why generic advice tends to make half its audience worse.

The first is the obvious one: the turn starts on the bounce instead of on the opponent's contact. Everything downstream compresses, the swing gets rushed into a shorter path, and contact retreats toward the hip. This is the error the reaching instruction is trying to solve.

The second is stranger and less discussed. The player turns early — genuinely early, good preparation — and then carries the racquet too far back, or stops and waits in a static coiled position. Both cost time. A takeback 40 centimeters longer than necessary, at a mean forward hand speed of around 12 m/s, adds roughly 33 milliseconds to the swing. Look back at the table: that's most of a racquet head's worth of contact-point drift, bought with what felt like good preparation. And a swing restarted from a dead stop has to be launched on a guess, because the continuous steering Bootsma described works better when the system is already in motion.

Early and stalled fails in the same place as late and rushed. The two players will describe identical symptoms.

What actually changes between Tuesday and Thursday

Here is the part that should be reassuring, and it isn't self-help — it's a claim about inputs.

The ball varies more than the player does. The ITF specification for a standard ball permits a rebound between 135 and 147 centimeters when dropped from 254 centimeters onto concrete. That is a legal band of about 12 centimeters, roughly eight percent, and that's before a ball has been hit forty times and gone fluffy. The ITF's Court Pace Rating sorts surfaces into categories from slow to fast based on how much horizontal speed the court takes out of a ball. Add incoming spin, which changes both bounce angle and post-bounce speed, plus the depth of the shot you're receiving, plus wind, and the arrival time of the ball at your hitting zone is a distribution, not a constant.

Against an eleven-millisecond window, small shifts in that distribution are enough to move a player from clean to cramped without a single thing changing in their swing. This is not a reason to stop working on technique. It is a reason to stop reading day-to-day variance as evidence about your technique.

Three checkpoints and a way to check them

The checkpoints below are landmarks, not laws. They are useful because they replace an unmeasurable instruction ("time it better") with observable events that occur far enough upstream to be controllable.

• Shoulders and racquet turned before the ball crosses the net. Not before it bounces — before it crosses. This is roughly 400 to 600 milliseconds of margin on a rally ball.
• Non-hitting hand across and pointing before the bounce. It functions as a shoulder-line indicator you can see in your peripheral vision, which is why coaches keep returning to it.
• Forward swing initiated at the bounce, not after it. The bounce is the anchor the visual system already prefers.

To verify: film from behind, in slow motion. A phone shooting 240 frames per second gives about 4 milliseconds of temporal resolution, which is genuinely enough — you're looking for errors of 25 milliseconds and up, which is six frames. Find the frame where the ball crosses the net and check the shoulders. Find the frame of the bounce and check whether the racquet has started forward. Two frames, two yes-or-no answers. That is the whole diagnostic, and it beats any amount of feel-based introspection, because feel is generated after the fact by a system that doesn't have millisecond access to its own operation.

One rule of thumb, stated as a directive because the evidence supports this one: start the turn on your opponent's contact, and start the forward swing on the bounce. Both events are external, both are visible, and both sit far enough upstream that a conscious decision can still reach them.

What's solid, what's thin, what's folklore

Well-established. Impact duration in the low single-digit milliseconds. The proximal-to-distal sequencing of the forehand and its association with skill level. That skilled players extract anticipatory information from pre-contact opponent kinematics — this one has been replicated across sports and paradigms. And the arithmetic linking timing error to contact-point displacement, which isn't a finding so much as a consequence.

Plausible but thin. The specific gaze prescriptions. Damien Lafont's preliminary photographic analysis in the International Journal of Performance Analysis in Sport, examining head and gaze behavior through the hitting phase, suggested that elite players hold the head still through contact rather than tracking the ball onto the strings — but it examined a handful of players from photographic sequences. The non-dominant hand as a timing lever is coaching consensus with, as far as we can find, very little direct measurement behind it. It is probably right. It has not really been tested.

Folk wisdom. "Hit it further out in front" as a remedy. "You were late because you weren't focused." Both describe the outcome and mistake it for a cause.

The question we can't close

Bootsma's table tennis players were correcting inside the stroke, and correcting well — their timing at contact was tighter than the movement that generated it. If that machinery is as good in a 3.5 club player as it appears to be in an expert, then a bad day may not be a swing that broke. It may be a swing whose corrections ran out of runway, which is a completely different problem with a completely different fix.

But nobody, as far as we could find in the biomechanics and motor control literature, has measured the distribution of contact points for the same recreational player across weeks — good days and bad days, same person, same camera. Everything above is inferred from lab kinematics on small expert samples and from arithmetic. The inference is reasonable. It is still an inference.

So the open question isn't whether timing beats technique. It's this: when a 3.5 player's forehand deserts them for a fortnight, has the movement itself changed, or has only its arrival time changed — and if it's only arrival time, why would grooving the movement fix anything?