Double Poling Power Test: Pole Impulse and Left-Right Asymmetry Protocol for Cross-Country Skiers

A junior national team coach pulls up the season's double poling test sheet and everything looks like it should: watts per kilogram on the roller-ski ergometer keep climbing, the 1000m double-poling-only time trial keeps dropping, threshold heart rate sits right where the plan wants it. Then the team physio mentions soreness on one side of the lower back and lat, always after a heavy double poling block, never during classic sessions with a kick. Nothing on the test sheet explains it, because nothing on it was ever built to separate what the left pole did from what the right pole did.

Double poling is one of the few endurance tests where nearly the entire propulsive job routes through two brief, simultaneous ground contacts per cycle, roughly a third of a second apiece at race pace. A skier can post an excellent average power number while quietly loading one side more than the other, because the ergometer, the stopwatch, and a coach watching from the sideline only ever see the combined result. The protocol below pulls that sum apart on roller skis with a wearable IMU setup, measuring impulse per pole plant and building a left-right symmetry index a season average can't surface.

Peak power and average speed both answer how much total propulsive work got done. Neither answers what a coach managing an overuse issue needs to know: whether that work got shared evenly between two arms doing structurally the same job at the same moment. Because both poles plant together rather than alternating the way a running stride does, a skier can compensate for a weaker or restricted side by loading the stronger side harder within the same ground-contact window, and the combined force or velocity trace looks completely normal from the outside. That compensation shows up as one-sided soreness months before it shows up as a slower time; the only way to catch it early is measuring both sides separately, on the same trial, at the same instant.

A double poling cycle breaks into a poling phase, pole plant to release, when the athlete generates essentially all forward propulsion, and a recovery phase, the swing back overhead, generating none. Poling time (Tp) divided by cycle time (Tc) gives the duty factor, and Holmberg, Lindinger, Stöggl, Eitzlmair, and Müller (2005), instrumenting elite skiers' poles with strain-gauge transducers on a roller-ski treadmill, found duty factor shrinks as speed rises: faster skiers generate more force and impulse in a shorter window, not by pushing longer. Separating left-pole impulse from right-pole impulse needs either two independently instrumented poles or the field workaround built around per-pole sensors below.

Wind wrecks this test outdoors more than anything else. A 15-20km/h headwind or tailwind shifts poling time and duty factor independent of the athlete's arms, so treat a breezy outdoor session as a rough technique check only. A treadmill or a still, sheltered road on a calm morning removes that variable entirely.

1. Warm-up (10-12 minutes): easy diagonal-stride rollerskiing and shoulder/thoracic mobility work, then 3 progressive double-poling-only passes building from roughly 70% to 90% effort.
2. Sensor check: zero both pole IMUs and the torso IMU with the athlete standing still for 3 seconds, then run 3 exaggerated, deliberately symmetric pole plants at a standstill to confirm both pole sensors fire on the same events.
3. Steady-state stage: 60-90 seconds of double-poling-only at a fixed submaximal pace, either a constant treadmill belt speed around 3.0-3.5 m/s at a 1-2% incline, or a controlled flat-road pace held with even cadence. Scored for symmetry, not capacity.
4. Recovery: 3-4 minutes of easy skiing or standing rest.
5. Maximal stage: one maximal double-poling-only effort, a 100m flat-road sprint from a rolling start or a 20-25 second push on the treadmill. Scored for capacity, not symmetry, since asymmetry can spike under fatigue in ways that don't reflect the baseline pattern.
6. Valid cycle criteria: discard any cycle where the pole IMUs register plant timing more than 40-50ms apart; a larger gap usually means one pole skipped or skidded rather than planted cleanly.
7. Scoring: average at least 8-10 clean cycles from the steady-state stage for symmetry, and the mean of the 3 highest-impulse cycles from the maximal stage for capacity.

Total time per athlete, warm-up included, runs about 20-22 minutes. Run both stages the same session; splitting them across days lets day-to-day fatigue swamp any real change in symmetry.

Estimating impulse without a force-plated pole starts from Newton's second law rather than a pressure sensor: force equals mass times acceleration, so impulse equals mass times the change in velocity it produces. During the poling phase, the torso IMU's velocity trace rises as the poles drive the athlete forward; that rise, multiplied by body mass, is the total propulsive impulse from both poles on that plant: J_total = m x delta-v(poling phase).

Splitting that number between the two poles is where the per-pole IMUs earn their place. Each pole sensor records its own peak loading signature, the deceleration spike as the tip bites into the ground, and the relative size of the left and right spikes gives a share of the combined impulse: Impulse_left = (peak_left / (peak_left + peak_right)) x J_total, and the same for the right side. Not a direct force-plate reading, but consistent enough to track one athlete's own left-right balance over time.

Symmetry is expressed as a Double Poling Asymmetry Index (DPAI): DPAI = |Impulse_left - Impulse_right| / ((Impulse_left + Impulse_right) / 2) x 100. Worked example: a 72kg skier's torso IMU shows a 0.55 m/s velocity increase across a 0.34-second poling phase, giving a combined impulse of 72 x 0.55 = 39.6 N-s. The pole sensors show a 55/45 split, left over right: 21.8 N-s and 17.8 N-s. DPAI works out to roughly 20%, worth flagging even though cadence, speed, and overall power looked unremarkable on their own.

Two studies get cited most often here, and it's worth being precise about what each actually measured.

Holmberg, Lindinger, Stöggl, Eitzlmair, and Müller (2005), in Medicine & Science in Sports & Exercise, instrumented poles with in-shaft strain-gauge transducers and tested 9 elite male skiers double poling on a roller-ski treadmill across a range of speeds. Peak pole force and impulse both rose strongly with speed, correlations exceeding r = 0.85, while poling time and duty factor shrank as speed increased: faster skiers generated more force and impulse in a shorter window, not by pushing longer. Limitation: 9 elite male athletes at one controlled incline, on instrumented poles adding roughly 100g of swing weight versus a race pole; the findings describe elite technique and were never meant as benchmarks for recreational skiers.

Onasch, Killick, and Herzog (2018), in Sports, independently instrumented the left and right poles of skiers double poling on a treadmill to compare dominant-side and non-dominant-side peak force within the same cycle. Measurable within-subject asymmetry showed up in most skiers, with no consistent direction across the group; self-reported hand dominance didn't reliably predict the stronger side, and asymmetry ranged from negligible to differences the authors flagged as practically meaningful. Limitation: a modest sample at one fixed treadmill speed and incline, so it doesn't establish how asymmetry behaves under fatigue, race pace, or overground skiing.

Together, the two studies support what the protocol does: track force and impulse against speed the way Holmberg's group demonstrated, and treat asymmetry as an individual, trending measure rather than a fixed norm, since Onasch's group found no reliable population-level pattern to compare a skier against.

Peak pole force and duty factor from the maximal stage are field-practical bands, not a strict cutoff, since a wearable setup carries more noise than lab-instrumented poles.

A longer relative poling time usually reflects slow force application, not low effort; a strong, fit skier can still show a long duty factor if the poling motion itself is slow to load, a technique target distinct from raw strength.

The bands below apply to the steady-state stage's DPAI, since that stage exposes resting technique rather than fatigue-driven breakdown.

Compare a maximal-stage DPAI only against other maximal-stage sessions, and a steady-state DPAI only against other steady-state sessions; fatigue reliably pushes asymmetry higher regardless of the underlying pattern, so mixing the two makes a normal maximal effort look like a technique problem.

An elevated DPAI is a training target, not a verdict. Skiers in the notable or significant bands typically respond to two things done together: unilateral strength work on the weaker side, tempo pull-downs and single-arm cable rows loaded toward the underperforming pole, and a video cue focused purely on symmetric plant timing rather than power, since chasing more force on the weak side without fixing timing usually just builds a stronger asymmetric pattern.

Retest every 4-6 weeks rather than every session; DPAI moves slower than sprint or time-trial numbers, and testing too often mistakes noise for real change. If a physio flags one-sided soreness, retest before returning to full volume, since a DPAI still elevated relative to baseline is a better readiness signal than how the shoulder feels that morning. If speed and power look excellent while DPAI keeps climbing, treat that as the early warning it is, months before it shows up as a slower time.