In hockey, most competitive situations are decided within the first three to five strides. The race to a loose puck, the gap created off the rush, the recovery step that puts a defender back between the puck and the net — these are settled before the skater has reached anything close to maximum velocity. What happens in those strides, how they are built, and whether the testing and training systems designed to develop them are actually targeting the right thing: that is what this article is about.

Good numbers. Wrong conclusion. That combination may be the most expensive diagnostic error in hockey performance — and it is happening in weight rooms and testing sessions across every league in Europe right now.

There is a belief embedded in the culture of hockey performance that is rarely questioned and regularly costs athletes development time. It goes something like this: explosive power is explosive power, speed is speed, and a player who jumps far and squats heavy will accelerate faster on the ice. The logic feels clean. The data seems to confirm it. And it is quietly leading coaches to the wrong conclusions.

The problem is not the testing. Broad jump, squat jump, countermovement jump, and drop jump are legitimate tools with real predictive value in the right context. The problem is the interpretation — specifically, the habit of borrowing the interpretive framework from sprint athletics and applying it to a movement that is mechanically, directionally, and energetically different in almost every meaningful way.

Acceleration on Ice Is a Different Problem

When a 100-metre sprinter accelerates off the blocks, the dominant mechanical demand is horizontal force production against a stable, high-friction surface, repeated in a single plane over a relatively long time frame. The athlete leans forward, extends aggressively, and cycles through a fairly predictable force-time curve. That pattern is what broad jump, squat jump, and sprint-based testing protocols are built to reflect.

Hockey acceleration does not look like that.

The first three to five strides out of a standing or skating start involve forward lean combined with substantially lateral force application — the push directed well away from the line of travel rather than along it — short contact phases through a blade edge rather than a flat foot, and an immediate transition into a glide phase where force application stops entirely between pushes. Research conducted on ice at the University of Calgary's Olympic Oval, using accelerometry, electromyography, goniometers, and in-skate plantar force sensors during 30-metre skating trials, found that acceleration strides involve significantly greater plantar-flexor muscle activity, greater hip-extension emphasis, and larger plantar push-off forces than steady-state strides. Separately, high-calibre players demonstrated greater hip range of motion and greater forefoot force application than lower-calibre players across both stride types. Critically, the calibre differences between groups were most prominent during the acceleration phase — high-calibre players completed the first ten metres 15% faster than lower-calibre players, but only 7% faster in steady-state. Lower-calibre players also showed a significantly larger performance gap between their acceleration and steady-state times, meaning technique breaks down earlier and more severely precisely when the demand is hardest (Buckeridge et al., 2015, PLOS One).

The ice surface introduces friction coefficients that depend on blade angle, edge sharpness, and load distribution — variables that have no equivalent in overground sprinting. Skating stride force is blade-mediated. What reaches the ice is filtered through the angle of the skate, the quality of the edge, and the athlete's ability to time their push relative to the glide. Two players with identical CMJ scores can look entirely different in their first five metres on ice — not because one is stronger, but because one directs their capacity into the skating pattern and one does not.

What the Tests Actually Measure — and What They Don't

This is not an argument against off-ice testing. It is an argument for precision in what coaches claim those tests tell them.

Broad jump measures horizontal power output in a single maximal effort from a bilateral stance. Squat jump isolates concentric lower-body power without the elastic energy contribution of a countermovement. CMJ adds the stretch-shortening cycle and reflects reactive capacity and neuromuscular readiness. Drop jump targets reactive stiffness and landing mechanics under high-load conditions.

Each of these tests explains some portion of skating speed variance. The research supports that. But the strength of those relationships varies considerably depending on player level and which skating quality is being measured. A 2024 study of U16 elite and sub-elite players found that for elite players specifically, the CMJ correlated with short skating acceleration at only r = −0.46 and the broad jump at r = −0.32 — both firmly in the moderate range. Sub-elite players in the same study showed considerably stronger relationships between off-ice power and sprint performance. The authors conclude directly that elite players' skating sprint is less related to their vertical and horizontal take-off abilities than in sub-elite players — and that sub-elite players appear to compensate for less refined technique by relying more heavily on raw strength and power, a strategy that generates greater fatigue in repeated sprint efforts (Roczniok et al., 2024, Journal of Human Kinetics). The skating efficiency index — a direct measure of stride technique and glide quality — correlated moderately with both short and long sprint performance in elite players, suggesting that at the highest level, what separates fast from slow skaters is less about force production and more about how effectively that force is converted into forward motion through the skating pattern.

This distinction between player levels matters for a reason that is easily missed. For younger and less developed athletes, general athletic qualities — strength, power, basic speed — are genuinely more predictive of on-ice performance because skating skill has not yet become the dominant variable. That is not a flaw in the research; it is an accurate description of a developmental reality. The argument here is not that general athletic development is irrelevant. It is that as players mature and skating technique becomes more refined, the predictive relationship between off-ice power and on-ice acceleration weakens — and the training system needs to reflect that shift, not ignore it. A testing battery appropriate for a sub-elite programme should not be driving decisions in an elite one.

Many coaches have seen a version of this in the field: a player retests better on a standing long jump after a simple mobility or activation intervention — hip flexor stretching, a glute and hip extensor reset — without becoming any stronger in the physiological sense. That does not make the intervention magic. It simply confirms that a jump score is a snapshot of expressed readiness under specific conditions on a specific day, not a fixed verdict on skating potential. The numbers carry value. But the moment a coach treats a jump score as a verdict rather than a data point, the system starts to mislead.

The First Strides Are Their Own Problem

The opening strides of a hockey acceleration deserve separate attention, because they are where the greatest mismatches between off-ice testing and on-ice outcome actually live.

Coming out of a stationary start, the skater must generate lateral force while simultaneously leaning forward, manage a short contact phase through a narrow blade edge, and begin the transition to glide before the push is fully complete. There is no equivalent to this movement pattern in a broad jump or squat jump. The direction of force, the contact geometry, the timing of the push relative to glide onset — all of it diverges from what any jump protocol requires.

The game-context question that rarely gets asked is this: how often does a player actually produce a clean, maximal acceleration in a competitive game? Time-motion analysis of elite international hockey shows that forward sprinting accounts for approximately 5% of effective playing time, within a broader high-intensity activity profile of around 18% of effective playing time. As the game progresses, the time spent in forward sprinting in the third period declines by 54.8% compared to the first — a significant, statistically confirmed drop that signals genuine fatigue-related impairment of sprint capacity as the match develops (Brocherie et al., 2018, Biology of Sport). Most in-game accelerations are short, contested, and interrupted within the first five to eight metres by a check, a puck decision, or a change of direction. The strides that win the situation — the ones that matter most in a competitive game — are almost always over before the player has reached steady-state velocity. Clean, uncontested maximal acceleration runs are a rare event, not the recurring demand the training model often assumes.

That context reframes the development question considerably. If the majority of in-game accelerations are short, reactive, and technically demanding, then the marginal return on improving skating mechanics and first-stride technique in technique-limited athletes may substantially exceed the return on another block of general strength or power training. Capacity is not the constraining factor for every player. For some, it is not even close to the main one.

What Does the 10-Metre Skate Actually Look Like?

Timing splits are necessary. They are not sufficient.

A player who covers ten metres in a competitive time can still be leaving significant speed on the ice through poor first-stride mechanics — an upright trunk in strides one and two, insufficient lateral push angle, early edge rollover, or a glide phase that bleeds momentum before the next push lands. The clock gives you the outcome. It does not give you the mechanism.

This is why watching how the 10-metre skate looks is not optional qualitative work layered on top of the real data. It is the real data. The stride pattern in the first three pushes, the degree of forward lean, the width and angle of the first lateral extension, the timing of the transition from push to glide — these are the variables that determine whether a player with a strong physical base is converting that base into skating speed or leaving it behind in the gym. A coach who only reads the time and moves on is stopping exactly where the diagnosis should begin.

Squat Strength and Jump Power Are Not Skating Speed

Increased squat strength and improved jump metrics raise an athlete's ceiling. They expand what is possible. That is genuinely valuable, and coaches are right to develop it.

But realized stride force — what actually gets transferred into the ice on each push — is a function of how effectively the athlete's capacity is directed into skating technique. Strength and power training do not automatically improve that efficiency. They provide more raw material to work with. The conversion of that material into faster first strides is skating work, not gym work.

Research on vertical jump and on-ice performance in junior hockey players found that while drop jump correlated moderately with forward average skating speed over 30 metres (r = 0.62), CMJ showed no significant correlation with forward skating performance at all — its only significant skating correlation was with backward skating speed (Gupta et al., 2023, Biology of Sport). The drop jump finding is worth reading carefully: even the best-performing off-ice test correlates only moderately with a measure that includes a substantial non-acceleration portion of the sprint distance. For truly short acceleration — the first five metres where situations are won and lost — no off-ice jump test in the current literature demonstrates a relationship strong enough to substitute for measuring the thing directly on ice.

The Danger of One-Speed Thinking

The deepest problem here is conceptual. When coaches import the logic of sprint athletics into hockey — broad jump predicts sprinting acceleration, skating is acceleration, therefore broad jump predicts skating acceleration — they construct a framework that is plausible on the surface and systematically flawed underneath.

The consequences are practical and consistent. Jump scores are weighted too heavily when evaluating or ranking players for skating speed. Off-ice work is over-prioritised in development programmes at the expense of on-ice acceleration mechanics, edge work, and start-specific skating skill. Players who score well on generic athletic benchmarks get credited with skating qualities they have not demonstrated on ice, while players with moderate jump scores and excellent skating mechanics are consistently undervalued.

None of this is careless. It is the entirely predictable result of a testing vocabulary borrowed from a different sport and applied without sufficient interrogation. And the numbers themselves are not the problem. A below-average broad jump score tells you something meaningful about horizontal power expression. It tells you nothing about whether the underlying issue is force production, blade mechanics, lean angle, push timing, or expressed readiness on test day. Off-ice data describes capacity. It does not diagnose skating failure. And critically, it cannot tell you whether the player in front of you is force-limited or technique-limited — the single most important diagnostic question in individual player development, and the one that should be driving every training decision that follows.

A Hierarchy That Reflects the Actual Problem

The practical correction is not to abandon off-ice testing. It is to reorder the hierarchy and be rigorous about what each layer of data is actually answering.

On-ice testing — 5, 10, and 30-metre timed starts, repeated skating sprint protocols, and skating-specific agility assessments — should serve as the primary data source for evaluating hockey acceleration. These tests measure what matters, in the environment where it matters. And they should always be accompanied by direct observation of how the skate looks, not just how long it takes.

Off-ice jump tests occupy a secondary role: they indicate the power base and athletic capacity an athlete brings to the skating pattern. They are descriptors of potential, not predictors of on-ice outcome, and they are not diagnostics of skating acceleration failure.

Strength and power training should be understood as supportive infrastructure. The goal is to build the physical qualities that skating mechanics can then express. Where a player is genuinely force-limited, more strength work is the right answer. Where a player is technique-limited, more strength work is the wrong answer delivered with confidence.

There is a practical objection worth addressing directly, because it will occur to any experienced coach reading this. Off-ice testing is controllable, repeatable, and easy to deliver at scale. A force plate, a timing gate, a jump mat — these are available, reliable, and produce numbers that feel authoritative. On-ice acceleration assessment requires ice time, skilled observation, and the kind of qualitative reading of stride mechanics that takes years to develop. The difficulty of doing on-ice assessment well is real, and it explains why off-ice testing has filled the gap. But difficulty of measurement cannot determine the hierarchy of what matters. If the primary variable is skating acceleration and the available proxy is a gym-based jump test, the solution is to invest in better on-ice assessment — not to elevate the proxy to a primary position because it is more convenient to administer. Convenience is not validity.

The Question Most Programmes Are Not Asking

Let me be clear about something before going further. I am a gym person. I believe in strength, I believe in power, I believe in what a well-designed training programme can build in an athlete over time. The gym is not the problem. My own bias toward it might be — and that is exactly the point. As a coach, I cannot allow what I love about training to override what a specific player actually needs. The moment personal conviction starts answering the diagnostic question, the player pays the price. What follows is not an argument against strength work. It is an argument for honesty about when strength work is the answer and when it is not.

Most programmes measure preparation in hours. Hours in the gym, hours on the ice, hours in the conditioning block. It is a metric that is easy to report, easy to defend, and almost entirely silent on what actually matters — whether those hours are solving the right problem for the right player.

The base must come first. Sufficient strength, power, and force capacity are not optional; they are the physical floor beneath everything else. Without them, there is nothing to express on the ice. That is settled and non-negotiable.

What is worth questioning is what happens after that floor is built — and whether the hours that follow are directed at the actual constraint or simply added to a programme that already has enough of what it is already good at.

For a meaningful number of players, the constraint is not force output. It is what happens to that force between the gym and the ice edge. A player with unresolved first-stride mechanics, a shallow push angle, or poor transition movement is not undertrained. He is misdirected. And the answer to that is not another training block. It is purposeful time, one-on-one, with a skilled skating coach — on the ice, in the environment where the problem actually lives — working through the specific movement demands that player faces in a real game. Not a linear acceleration template. The pivots, the board-battle exits, the tight-angle changes of direction that determine whether he wins or loses the moments that decide games.

The cost of ignoring this is higher than most programmes acknowledge. Research on elite versus sub-elite youth players confirms it directly: players who rely on strength and power without refined skating technique show greater energy expenditure and significantly greater fatigue during repeated sprint efforts (Roczniok et al., 2024, Journal of Human Kinetics). Strength and power added onto unresolved skating mechanics do not produce a faster player. They produce a greater workload moving through an inefficient pattern, stride after stride, shift after shift — which is precisely why that player looks progressively worse in the third period, not better. More capacity into a flawed system does not amplify the solution. It amplifies the flaw.

The question worth sitting with this summer is not how many hours are on the programme. It is whether the hours already there are pointed at what is actually missing — and whether the player standing in front of you needs the gym, or needs the ice, and someone on it who can finally show him why his first stride is costing him the game. That is a conversation worth having, and one I will return to in a separate piece.

Practical Takeaway

Hockey acceleration and sprint acceleration share a physical foundation but differ in direction, surface, contact mechanics, and movement pattern in ways that make direct translation unreliable. Off-ice jump and strength data describe capacity — they do not diagnose skating failure, and they cannot tell you why a player accelerates poorly on ice. On-ice timed starts combined with direct observation of skating mechanics should anchor your evaluation hierarchy, with off-ice tests informing the picture rather than defining it. The most important diagnostic question in hockey performance is not how strong or how powerful a player is — it is whether the physical qualities he has built are actually reaching the ice.

References

Brocherie, F., Girard, O., & Millet, G.P. (2018). Updated analysis of changes in locomotor activities across periods in an international ice hockey game. Biology of Sport, 35(3), 261–267.

Buckeridge, E., LeVangie, M.C., Stetter, B., Nigg, S.R., & Nigg, B.M. (2015). An on-ice measurement approach to analyse the biomechanics of ice hockey skating. PLOS One, 10(5), e0127324.

Gupta, S., Baron, J., Bieniec, A., Swinarew, A., & Stanula, A. (2023). Relationship between vertical jump tests and ice skating performance in junior Polish ice hockey players. Biology of Sport, 40(1), 225–232.

Roczniok, R., Stastny, P., Novak, D., Opath, L., Terbalyan, A., & Musalek, M. (2024). The relation of on-ice and off-ice performance at two different performance levels in youth ice-hockey players. Journal of Human Kinetics, 93, 193–203.

Magnus Ågren

Performance and Leadership Development · Consultant · SHL - NL - DEL

Thirty years in elite sport. Seven seasons as Head of Performance and Medical in the Swedish Hockey League. Olympic cycles since Sydney 2000. Designs the systems that integrate coaching, medical, and sports science into one performance structure.

www.magnusagren.com

People. Purpose. Performance.