The Shift Repeatability Protocol: Testing Whether a Hockey Player's Power Holds Across Every Shift

Every bench has a player like this: first two shifts look sharp, wins the race to the puck, finishes the check, gets back to the bench inside the 45-second window the coach wants. By the fourth or fifth shift of the second period, the same player still wins that race - just half a stride slower, and that half-stride is the difference between a clean breakout and an icing call. Coaches notice it by eye long before anyone runs a number on it, and the usual response is to shorten that player's shifts or shuffle the line. That fixes the symptom for one game. It doesn't say whether the player has an actual conditioning gap or just had a rough night.

The physical demand hockey places on a skater isn't a single all-out sprint, and it isn't steady aerobic work either - it's 40 to 60 seconds of near-maximal, high-speed skating punctuated by 90 to 120 seconds on the bench, repeated somewhere between 15 and 25 times over three periods depending on line deployment. That rhythm - short brutal effort, short incomplete rest, repeat for an hour plus - is what a shift repeatability test measures, and it's a different question from the one a Wingate or a standard repeated-sprint test answers. This guide lays out a six-rep protocol built around that exact work-to-rest ratio, what the research says about where the demand actually sits, and how to read the resulting fatigue number without turning one noisy session into a roster decision.

A Wingate test asks for one 30-second all-out effort against a fixed resistance, then it's over - no second bout, so it can't tell you anything about output on the fifteenth repeated maximal effort of the night. RAST and most repeated-sprint protocols get closer, but they're typically built around six-second-range sprints with 10 seconds of active recovery, which loads the phosphocreatine system in short bursts and recovers it almost fully between reps. A hockey shift is neither of those things: it runs four to ten times longer than a single RAST sprint, and the rest that follows it - stepping over the boards, sitting on a cold bench, waiting for the next change - is shorter relative to the work than a lab protocol would ever schedule on purpose.

That mismatch matters because the system doing most of the work shifts as effort duration stretches past about 30 seconds. Phosphocreatine dominates the first 10-15 seconds of a shift, but glycolytic and oxidative pathways take over an increasing share from there, and incomplete bench rest means a player starts the next shift without having cleared that fatigue. A test built around 6-second bursts and near-full recovery never puts an athlete in that state, which is exactly the state a coach cares about at 14 minutes into the third period. Testing the actual shift rhythm - not a proxy for it - is the only way to see whether a player's engine holds up under the specific stress the position creates, a gap we also address from the skating-power side in our guide to building explosive power for hockey.

The Shift Repeatability Test (SRT-6) asks an athlete to complete six maximal-effort bouts of 45 seconds each, separated by 105 seconds of passive bench-style recovery - a roughly 1:2.3 work-to-rest ratio mirroring what time-motion research has measured during games rather than a rest window picked for lab convenience. On ice, the effort is a repeated cone-to-cone skating shuttle (forward stride, tight turn, forward stride, for the full 45 seconds) covering as much distance as possible; off-ice, an assault bike, ski-erg, or slide board at maximal cadence substitutes reasonably well for summer testing or programs without ice access, provided the same modality is used for every retest.

Distance covered (on ice) or average power (off-ice) is logged for every rep, along with heart rate at the end of each work bout and again at the end of each recovery window, plus a 0-10 RPE taken immediately after rep six. Athletes should be told explicitly to treat rep one as an all-out maximal effort rather than pacing themselves - a sandbagged first rep is the single most common way this test gets corrupted, since it makes every later rep look artificially close to rep one and hides the decay the protocol exists to detect.

The 1:2.3-ish work-to-rest structure in the SRT-6 isn't an arbitrary round number - it comes from one of the foundational time-motion studies in the sport. Green, Bishop, Houston, McKillop, Norman, and Stothart (1976, Journal of Applied Physiology) filmed and physiologically monitored university-level ice hockey players during competition and found average shift durations of roughly 40 seconds, ranging from about 30 to 85 seconds depending on position and situation, with heart rates during shifts reaching 87-95% of measured maximum and post-shift blood lactate commonly sitting in the 4-8 mmol/L range. That combination - sustained near-maximal heart rate plus meaningfully elevated lactate on a 30-85 second effort - is the physiological signature the SRT-6 tries to reproduce on demand rather than wait to observe live. The limitation worth naming: the sample was small, male, and collected in the 1970s under a slower rule set than today's game; several current staffs target shifts closer to 35-45 seconds given current tempo, so treat the study's numbers as a directionally reliable floor rather than a precise modern target.

The second anchor is Glaister's 2005 review in Sports Medicine synthesizing repeated-sprint-ability research across sports and protocols. Glaister reported that trained team-sport athletes typically show a 5-10% decrement across a repeated-sprint series, with well-conditioned athletes clustering toward the lower end and decrements above 10% associated with reduced repeated-effort capacity and, in several reviewed studies, lower aerobic fitness markers. That range is a useful reference point for an SRT-6 score, though the review's underlying data comes overwhelmingly from short (4-10 second) running sprints rather than 45-second skating efforts, so treat the cutoffs as a starting reference to calibrate against your own squad's history rather than a hard line imported from a different sport and effort duration.

The primary output of the SRT-6 is the Shift Fatigue Index (SFI), calculated the same way a Wingate or RAST fatigue index is: (Rep 1 output - Rep 6 output) / Rep 1 output x 100. A skater who covers 62 meters on rep one and 54 meters on rep six posts an SFI of roughly 12.9%, which lands in the middle band below.

Treat a single session's SFI as a data point, not a verdict. Sleep debt, a heavy practice the day before, dehydration, and simple motivation all move the number by several percentage points independent of true conditioning status, and rep one being run half-effort (see the sandbagging problem above) can make a genuinely tired athlete's SFI look artificially clean. The more useful signal comes from testing the same athlete on the same modality three or four times across a preseason block and watching the trend line, not from treating one Tuesday-afternoon number as diagnostic on its own.

The failure modes here are rarely dramatic - small procedural drifts that make one session incomparable to the next. Letting recovery run long because the group is chatting stretches a 105-second rest toward 130-140 seconds, which changes how pre-fatigued rep three or four actually is and inflates every athlete's apparent repeatability. Switching modality between tests - skating shuttle in October, assault bike in March because the rink was booked - breaks the comparison entirely, since a bike spares the lower back and hip flexors from the eccentric loading skating imposes and almost always produces a flatter, more forgiving curve.

• Standardize the recovery window with a visible countdown clock, not a coach's rough estimate of two minutes.
• Never mix on-ice and off-ice testing for the same athlete's longitudinal record - pick one modality per athlete and stay with it.
• Brief every athlete on pacing before rep one; a coached maximal effort from the first second is non-negotiable for the fatigue index to mean anything.
• Log surface and footwear/blade condition, since a freshly cut sheet of ice skates faster than one near the end of a session and will shift distance-based scores independent of fitness.

None of these controls require expensive equipment - they require the same discipline that makes any repeated-effort protocol trustworthy, applied consistently enough that a change in the SFI actually reflects a change in the athlete rather than a change in how the test was run.

A high SFI is a training prescription, not just a red flag. The instinct for a lot of staffs is to respond with more steady-state aerobic conditioning - longer bike sessions, more distance running - but that targets a different energy system than the one actually failing during a 45-second shift with 105 seconds of rest. What moves an SFI down is repeated-effort work that mirrors the real stimulus: interval sets built around 40-second work bouts at 100-110 second recovery, run for 6-10 reps, two to three times per week for a 4-6 week block, adding a rep rather than lengthening rest as the primary progression.

Rugby sevens programs face a structurally similar problem - repeated short maximal efforts against incomplete recovery across a long match - and the periodization approach in our rugby sevens repeated-effort conditioning guide translates well to hockey once work and rest durations are matched to shift length rather than phase length. Retesting the SRT-6 every 4-6 weeks through preseason, then monthly in-season, confirms the block is working rather than assuming it is because the athlete feels less tired in practice - subjective fatigue and measured output decay don't always move together, particularly for players who've learned to mask a bad shift with body language that reads as fine on the bench.