What 10 Years of RSI Has Taught Me.

I gave a talk in 2016 on lower extremity stiffness. I had read something like forty articles and spent close to a hundred hours pulling it together, and I told the room straight away that not all of what I was about to say was strictly research based. Some of it was what I suggested, some was what I inferred, and some was what I still questioned. Ten years on, that is still the honest place to stand. The research has moved, my own testing has piled up, and the tools have gotten better, but the thing I cared about then is the thing I care about now. Coaches chase results. Stiffness, read correctly, is one of the cleanest ways to see whether the results are coming.

This is the first piece in a three part series, and I want to walk the same arc I walked that day, why this matters, what it actually is, and how you measure and train it, updated with what the last ten years taught us. By the end you should know why stiffness earns the title of key performance indicator, why it is better understood as reactive strength than as a fixed trait, and how a switch mat in your gym turns the whole idea into a number you can act on Monday morning.

Let me start with why I bother.

Why stiffness earns the title of KPI

Ask a coach what they care about and you will not hear "leg stiffness." You will hear speed, quickness, agility, leaping ability, deceleration, the ability to change direction and not lose a step. We are practitioners, and we care about what shows up on the field. So the only reason stiffness deserves a seat at the table is if it tracks the qualities we are already chasing. It does.

Lower extremity stiffness is strongly related to sprint velocity (Bret et al., 2002). Improvements in stiffness improve vertical hopping and jumping, because springs are what set stride frequency in bouncing gaits (Farley et al., 1991). Stiffness is closely tied to both isometric and dynamic strength and to the force an athlete puts into the ground, and it is associated with muscle cross sectional area and pennation angle (Secomb et al., 2015). The ability to change direction at a high rate is closely related to lower body stiffness, and it differs between levels of competition, which is to say better athletes tend to be stiffer in the ways that matter (Pruyn et al., 2015). And stiffer landings and plants are more economic and transition faster, at the cost of higher peak ground reaction forces (DeVita and Skelly, 1992).

That last point is the whole game in one sentence, so hold onto it.

The ten years since have only sharpened the case. We used to talk about stiffness as one number that was good to have more of, and we now know that is too blunt. A 2024 systematic review treated the Reactive Strength Index, the field measure I will get to, as a genuine key performance indicator across athlete populations, and it found that the single quality that most separated high reactive athletes from low ones was relative strength (Southey et al., 2024). Read that again. The athletes who are reactive are the athletes who are strong for their bodyweight. That is not a coincidence, it is the mechanism, and it is the reason I framed this whole topic the way I did in 2016.

"Stiffness matters because it is the application of strength. It is reactive strength.

That is the spine of everything that follows. Strength is the capacity. Reactivity is the expression of that capacity against the floor, fast, in the window of a single ground contact. If you only remember one line from this piece, remember that one.

The two-sided ledger: stiffness is tuned, not maximized

Before I tell you to go make athletes stiffer, I owe you the other half of the picture, because it is the half most people skip.

Stiffness lives on both sides of the injury ledger. Too little of it is associated with soft tissue injuries (Granata et al., 2001). Too much of it is associated with bony injuries, the stress reactions and stress fractures that come from a leg that does not give (Butler, Crowell, and Davis, 2003). It is also significantly influenced by what happens above the leg. An abdominal and trunk strengthening program can change leg stiffness during hopping, so the core is part of the spring whether you think about it that way or not (Dupeyron et al., 2013).

So the coaching idea is not "maximize stiffness." It is tune it. A spring that is too soft loses energy and lets the joint travel into ranges it cannot control. A spring that is too rigid transmits every bit of the impact straight into the bone. The right amount is specific to the task and the athlete, and a big part of being good at this is knowing how stiff is stiff enough for the sport in front of you.

I want to be honest here, because this idea gets repeated more confidently than the evidence supports. You will see the stiffness and injury relationship drawn as a clean U shaped curve, with a safe valley in the middle and danger on both ends. The direction of that idea is well supported, but most of the data behind it is cross sectional or retrospective, not prospective. A 2022 review of prospective cohort studies on limb asymmetry and injury found real associations, single leg hop asymmetry at or above ten percent raised injury odds in female athletes, eccentric hamstring asymmetry at or above fifteen percent raised them too, but it found the picture highly inconsistent across tests and populations, with no universal threshold that holds everywhere (Guan et al., 2022). So I will say it plainly. Do not tell an athlete that a specific stiffness value will hurt them. The relationship is nonlinear and still under investigation. Tune the spring, watch the athlete, and stay humble about exact cutoffs.

What stiffness actually is

Now the what. If we are going to train it, we have to agree on what it is.

Put simply, stiffness is the ability of a tissue or structure to resist deformation when force is applied. The physics is Hooke's Law (Butler et al., 2003; Alexander, 1997).

Force equals a constant, k, times the distance the object deforms, and that constant, k, is literally the stiffness. When something that obeys Hooke's Law deforms, its change in length is proportional to the force on it, and during that deformation it stores elastic energy, some of which is returned as the object springs back to its resting length. That is the entire reason we care. The leg is a spring that, loaded correctly, gives back much of what you put into it.

The spring is the muscle tendon complex. The classic mechanical model comes from Hill (1938) and frames the force the system produces as the summation of three parts. I teach it with everyday words, the way Zatsiorsky frames it in Science and Practice of Strength Training. The contractile element, the fascicles with their sarcomeres and cross bridges, is the machinery. The parallel elastic component, the sarcolemma and the muscle fascia, is the container. The series elastic component, the tendon and the aponeurosis, is the amplifier. Machinery, container, amplifier. That image has survived a decade of teaching because it is true enough to be useful.

Hardware and software

Here is the model I keep coming back to, the one that does the most work in my own coaching. Think of the muscle tendon complex as hardware and software. Structure versus function.

The hardware is structure. It is the cross sectional area of the muscle, its length, the fiber type, the amount of collagen in the tendon. It is the physical machine, and it changes slowly, over months and years. The software is function. It is motor unit recruitment, firing rate, synchronization, the activation and inhibition that happens with different movements and force levels, the co-contraction, the joint mechanics. It is how the machine is run, and it can change quickly, within a session in some cases.

The reason this matters is that the same hardware behaves completely differently depending on the software running on it. The exact same leg can be soft or rigid depending on the task and the moment, and that is not a flaw, it is the point. Stiffness is not a setting baked into the structure. It is something the nervous system dials in for the job at hand.

The counterintuitive part: muscle is compliant when passive, stiff when active

This is the slide that makes coaches sit up, and it is worth it. Muscle tissue is compliant when it is passive and stiff when it is active. The tendon is the opposite story, its stiffness stays roughly constant whether the system is loaded or not. So the springiness of the whole complex is not fixed, it is governed by how hard the muscle is working at that instant. In movement, the stiffer the active muscle, meaning the more the contractile element is contributing, the more the tendon can store and return as elastic energy. The degree to which the tendon stretches depends on the force put through it and the rate at which it is applied.

That reframes the way a lot of us talk about athletes. The "tight" athlete is not necessarily the stiff athlete, and the "loose" athlete is not necessarily the compliant one. Tightness at rest tells you very little about how the spring behaves under load. What matters is what the system does when it is asked to produce force fast, and that is a function thing, not a hardware thing.

How the spring works in a jump: the three jump continuum

If you want to see all of this in one place, line up three jumps. They sit on a continuum from no elastic contribution to heavy tendon contribution, the cleanest teaching tool I know.

The non countermovement squat jump starts from a paused, settled position. There is little to no contribution from stored elastic energy, because nothing was loaded on the way down. This is closest to pure contractile output.

The countermovement jump adds a dip and a reversal. Now there is more muscular compliance in play and a real but slower elastic contribution, with greater fascicle length change between the countermovement and takeoff. It is less dynamic than what comes next.

The drop jump asks the athlete to fall from a box, absorb the landing, and rebound. This is where the spring earns its keep. There is greater muscular stiffness and the most dynamic action of the three, and greater tendon length change between touchdown and takeoff.

And here is the number that is the most satisfying "aha" in the whole topic. In a drop jump, the tendons store about sixty six percent of the elastic energy and work output, and the fascicles account for about thirty four percent (Fukashiro, Hay, and Nagano, 2006). The fascicle contracts quasi isometrically, meaning it barely changes length, holding firm and producing high force, while the tendon does the lengthening and the recoiling through the amortization phase. The muscle holds the line so the tendon can do the springing. Two thirds of the rebound is the tendon. That single fact reorganized how I think about reactive training, because it means the muscle's job in a fast bounce is often to be a stiff anchor, not to shorten dramatically.

This maps onto a useful split in how we classify plyometrics. The stretch shortening cycle, the rapid stretch then shortening of the muscle tendon complex, comes in two flavors (Schmidtbleicher, 1992; Flanagan and Comyns, 2008). A fast stretch shortening cycle is under 250 milliseconds of ground contact, the pogos, tucks, bounds, and drops, with small joint angles. A slow one is over 250 milliseconds, the squat, broad, and max vertical jumps, with larger displacements that give more time to develop force. That 250 millisecond line is a memorable threshold that organizes every plyometric you will ever prescribe. Short contact and small angles is reactive, fast spring work. Longer contact and bigger angles is strength dominant power work.

How you build a spring: structure first, then function

This is the prescriptive heart of it, and it is where the hardware and software model pays off, because you build them in a specific order.

The structural pathway is the hardware, and here is the finding that should change how you sequence things. Tendon stiffness increases only after heavy load resistance training and isometric squats. Jumping, plyometric training, and sprinting produced no significant changes in tendon properties (Kubo et al., 2001, 2005, 2006, 2007). Let that sink in. The fast, springy, exciting work, the plyos, does not appear to build the tendon. Heavy strength does. The trainable increases in joint stiffness end up coming from changes in the muscle, not the tendon, but you need the heavy work to build and protect the tendon first.

So strength comes first, for two reasons (Turner and Jeffreys, 2010). It builds enough strength in the muscles and tendons to tolerate high plyometric loads without inviting tendon injury, and it builds the base the functional work then reshapes. You earn the right to chase reactivity by building the structure that can survive it.

The functional pathway is the software, and it is where plyometrics do their real job. When a muscle experiences sudden high tension, the Golgi tendon organ fires to inhibit the muscle, a protective brake that drops tension to keep the muscle and tendon from being damaged. After plyometric training, the inhibitory effect of that brake is reduced, so the athlete can sustain high landing forces with less exerted muscular force (Kyrolainen et al., 1991; Turner and Jeffreys, 2010). The plyometrics do not build the tendon, they teach the nervous system to stop protecting against the very forces you want the athlete to use, and they shift stiffness into the muscle so force gets expressed fast in short ground contacts. Structure buys the right, function spends it.

The newer work refines this without overturning it. A 2025 review and meta analysis of twenty three studies and 632 subjects found that plyometric or jump training produces a small but real effect on leg stiffness on its own, and the effect becomes large when it is combined with balance training or with resisted sprint training (Bandara et al., 2025). A 2021 study of twelve weeks of isolated plantar flexor plyometric training showed real structural change, with active muscle stiffness climbing and jump improvements tracking the mechanical changes rather than activation patterns (Kubo, Ikebukuro, and Yata, 2021). That was an isolated lab protocol, not a normal training block, so do not read its percentages as what your athletes will get, but the direction is clear. The spring is trainable, the structure does adapt, and it takes months to years, not weeks. In elite track and field jumpers monitored over four years, Achilles tendon stiffness ran about fifteen percent higher than in controls, and the takeoff leg was stiffer than the other (Karamanidis and Epro, 2020).

One more update belongs on the floor. Volume is not the lever we once thought. Across the lifespan, plyometric training improves the Reactive Strength Index, and programs running more than seven weeks at three sessions a week outperformed shorter ones, but similar gains showed up at fewer than 1,080 total jumps as at more than 1,080 (Ramirez-Campillo et al., 2023). More jumps did not buy more reactivity. In adolescents, countermovement jump improved best under 900 total ground contacts (Chen et al., 2023). Conservative volumes work, and they keep injury exposure down, a real change from the "more is better" instinct a lot of us grew up with.

"The fast work does not build the tendon. Heavy strength does. Earn the right to be reactive before you chase it.

How you measure it in the field

All of this would stay academic if you could not see it on your floor, so this is the part that turned the theory into something I use every day.

In a lab, human stiffness is calculated three ways (McMahon, Comfort, and Pearson, 2012). Vertical stiffness, Kvert, is peak vertical ground reaction force divided by peak vertical displacement, used for hopping and jumping. Leg stiffness, Kleg, accounts for the variables of running locomotion. Joint stiffness, Kjoint, is the change in joint moment divided by the change in joint angle. These are precise, and they need force plates and motion capture to compute honestly. I want to be square about that, because the rest of this piece rests on it. A switch mat does not give you a Kvert value, and I am not going to pretend it does.

What a mat gives you, cleanly, every rep, is two things: ground contact time and flight derived jump height. And from those two numbers you can reason about the spring with the field proxy that has held up for a decade, reactive strength.

The Reactive Strength Index is jump height divided by ground contact time, off a drop jump from a given height (Turner and Jeffreys, 2010).

It is the field expression of stretch shortening cycle ability, one number for how much height an athlete produces per unit of time on the floor. There are variants. RSImod uses jump height over the contraction time of a countermovement jump, so you can read reactive quality without a box. RSI-10 and the four jump pogo use average flight time over average contact time across a series of bounces. The pogo is my workhorse, because it shows what the athlete does rep after rep, not just on one heroic effort.

The reason ground contact time is so useful is that it is the most direct field read you have on how the spring is behaving. A short contact means the athlete reversed the load fast, which is what a stiff, well tuned spring does. A long contact means they sank into it and took longer to come back, which is what a more compliant system does. When you run RSI testing on a Plyomat, ground contact time is effectively a window into stiffness expression. You are not computing Kvert, but you are watching the behavior that Kvert is trying to capture, every single rep.

The research backs the practical value of this. Vertical stiffness is the most reliable of the lab measures and is more sensitive to high velocity running changes than leg stiffness, and validated field equations now let you estimate whole body stiffness without a force plate, which is exactly the direction practice has needed (Maloney and Fletcher, 2019). The Reactive Strength Index is no longer a lab curiosity, it is a validated performance indicator across populations, recognized as a composite of several stretch shortening cycle qualities rather than one trait (Southey et al., 2024).

Reading the shape: the Reactive Strength Quadrant

Here is where the measurement becomes a decision. If you plot ground contact time against jump height, every athlete falls somewhere on a two by two, and the quadrant they land in tells you what kind of reactive athlete you are looking at. In the Plyomat app these are the Reactive zone, the Compliant zone, the Stiff zone, and the Developing zone, drawn in their own colors, and the app places the athlete for you off their own reps.

Walk it in plain terms. High jump height and short contact is Reactive, the athlete who produces a lot in very little time, the prize. Long contact with good height is Compliant, the athlete who can jump but sinks to do it. Short contact with modest height is Stiff, quick off the floor but not yet converting that quickness into output. And low height with long contact is Developing, the athlete who needs the base built before reactive work makes sense. The quadrant is the diagnostic. It does not tell you what to program yet, that is a later post in this series, but it sorts the room in one test.

I should set expectations on one thing the way I set it for myself. You will see RSI norms quoted online, elite sprinters at three to four, recreational athletes around one to one and a half. Treat those as rough orientation, not gospel. The recent systematic reviews documented real differences between populations but did not publish a clean universal benchmark, and the commonly repeated numbers are hard to trace to a single primary source (Southey et al., 2024). Test your own athletes, build your own bands, and let the change over time tell the story rather than chasing somebody else's number.

Reading the depth, and reading each leg

Two newer pieces of the picture go beyond what I had in 2016, and they are where Plyomat's own work sits.

The first is depth. A single RSI describes the one box height you tested from, and nothing else. A short contact off a low box and a short contact off a high box give you the same number, even though the high box loaded the tissue far harder, so RSI cannot show you the depth at which an athlete stops rebounding and starts sinking. The honest fix is to test the same athlete across several drop heights and read the result as a curve, not a single setting. Where the curve peaks is the depth they are built to rebound from. Where it falls off is a training target.

This is the idea behind the Dynamic Rebound Index, a drop height aware reactive score developed by Lance Brooks. The work is his, and I think it is pointed at the right problem. It folds the eccentric loading of the drop, the total vertical displacement, into the math, so the score only improves when an athlete genuinely reverses a large downward momentum in a short time, rather than rewarding a quick foot off a shallow box. I want to be completely clear about its status, because intellectual honesty is the whole reason this topic is worth teaching. DRI is a conceptual extension, not a validated tool. As of now it has no published normative athlete data and no peer reviewed validation study with real athletes. It addresses a genuine blind spot in RSI and it is mechanically coherent with what the stretch shortening cycle does, but it is a framework, not settled science. Plyomat's incremental drop jump protocol, which samples RSI across several heights, is the practical, empirical version of the same idea. It shows you the curve without claiming a validated index.

The second is the single leg, and it is the read I would least want to test without. A two-leg RSI is a pooled, whole-system number: both legs and the trunk springing as one. It is useful, but one number cannot separate two legs, so it averages over the weaker side. Single-leg RSI resolves each leg on its own, and it is now established as the more sensitive read of lower-extremity reactive strength, particularly after injury. Two independent 2022 studies make the point starkly. In one, single leg drop jump RSI flagged functional deficits in athletes who had returned to sport after ACL reconstruction even when standard bilateral hop tests and balance tests showed nothing wrong (Lem et al., 2022). In the other, among ACL reconstruction athletes who had already passed the standard ninety percent limb symmetry hop test, only about thirty percent reached satisfactory RSI symmetry, meaning roughly seventy percent still carried a deficit the hop test missed (Hirohata et al., 2022). I keep meeting the same lesson on the rehab side, and I wrote up the return-to-play version of it separately, in the three single-leg tests I run to clear an athlete for sport, where a 2024 study found that roughly two-thirds of athletes who had passed their strength testing still failed reactive strength testing. The practical takeaway is direct. Measure each leg independently and compare them as a first class output, not an afterthought. The Plyomat app reads RSI per leg and surfaces the asymmetry so you see a side before your eye catches it. I will keep the caveat honest, because the prospective injury prediction value of any single asymmetry threshold is still inconsistent across the literature (Guan et al., 2022). The single leg read is for seeing what bilateral testing hides, not for promising who will get hurt.

What ten years of testing taught me

I will not close on other people's studies. I will close on what a decade of testing reactive athletes actually taught me, and I will say plainly that these are a coach's observations, earned on the floor, not a controlled trial.

A few things kept showing up, season after season. The athletes who changed direction best were almost never the ones who jumped the highest. They were the ones who spent the least time on the floor. For the sports that live on quick feet, contact time told me more than height ever did.

Stiffer was not always better. The times I chased maximal stiffness in an athlete who needed to absorb and redirect, I got brittle, not fast. The skill was never maximizing the spring. It was tuning it to the task in front of me.

Strength came first, every time. The athletes who could express real reactive strength were the ones who had built the base to tolerate it. When I let plyometrics outrun the strength work, the tissue is what paid for it.

And the single number always hid more than it showed. Two athletes would post the same Reactive Strength Index and need completely different training. That is where I stopped treating RSI as a grade and started treating it as a dial, a read on how much reactive load an athlete was ready for, rather than a score to chase.

These are patterns, not proofs. I offer them as what the model looked like on my floor, across a lot of athletes and a lot of seasons. The textbook gives you the mechanism. This part the textbook cannot give you.

The honest close

Stiffness is a key performance indicator, but it indicates differently depending on the sport and the athlete. A stiffer athlete tends to be strong in shorter to optimal ranges, a more compliant one in optimal to longer ranges, and a hybrid, the kind you find in a sprinter who has to be good across distances, in short, optimal, and long ranges alike. The coaching skill is not making everyone maximally stiff. It is matching the spring to the task, tuning rather than maximizing, because the same property that buys a fast shallow transition raises the peak force going through the bone.

And I will end where I started in 2016, with the caveat that gives the rest of it its weight. There is more to dynamic human movement than stiffness alone. It is necessary, it is trainable, it is measurable, and it is one of the most useful things you can put a number on. It is not the whole athlete. Treat it as a window, a clear and honest one, into how an athlete applies strength against the floor. When you run RSI testing on a Plyomat, ground contact time is that window, and the app reads the Reactive Strength Index, the per leg symmetry, the DRI style profile across heights, and the quadrant for you, so the shape of the athlete is obvious at a glance. The mat does not replace your eye or your judgment. It just makes the spring visible, every rep, so you can coach the athlete in front of you.

In Part 2 of this series I will take the single number apart properly, why two athletes can post the same RSI and train nothing alike, and how reading the shape across the quadrant changes what you prescribe. For now, the one thing to carry to the floor is the line we started with. Stiffness is the application of strength. It is reactive strength. Build the strength, then teach the athlete to spend it fast.

References