The Physiological Cost of Pacing Variability Coefficient: How Even Pacing and Negative Splitting Break the Marathon Sub-3 Barrier Through Energy Metabolism Dynamics

1. Introduction and Cutting-Edge Research Background

1.1 The Evolution from “Perceived Pacing” to “Scientific Pacing”

The evolution of marathon pacing strategies can be traced back to the 1950s, when Czech distance running legend Emil Zátopek captured gold in the 5,000m, 10,000m, and marathon at the Helsinki Olympics with his signature “violent surges.” However, Zátopek’s surges were fundamentally a tactic to psychologically pressure his opponents, not an optimal solution for energy efficiency. From the 1960s onward, as exercise physiology deepened its understanding of muscle glycogen depletion and lactate metabolism, coaches and sports scientists began to realize: pacing stability directly influences the efficiency of muscle fuel allocation.

Entering the 21st century, the proliferation of Global Positioning System (GPS) and running power meters allowed athletes to monitor their pace down to the second. In 2017, the International Association of Athletics Federations (IAAF) and the Japan Association of Athletics Federations (JAAF) collaborated to analyze split data from thousands of Tokyo Marathon finishers, uncovering a key phenomenon: runners finishing between 2 hours 30 minutes and 3 hours generally exhibited a Pacing Variability (PV) below 4%; while runners finishing over 4 hours often had a PV above 8%. This observation gave rise to the “Pacing Variability Coefficient” as a new paradigm for predicting marathon performance.

1.2 Definition and Calculation of the Pacing Variability Coefficient

The Coefficient of Variation of Pacing (PV) is calculated as the standard deviation of segment paces (typically per 5 kilometers) divided by the average pace, multiplied by 100%:

[
PV(%) = \frac{\sqrt{\frac{1}{n-1}\sum_{i=1}^{n}(v_i - \bar{v})^2}}{\bar{v}} \times 100
]

Where (v_i) is the average speed (km/h) of the (i)-th 5km segment, (\bar{v}) is the overall average speed, and (n) is the number of segments (typically 8 segments plus the final 2.195km for a full marathon). A lower PV value indicates more uniform pacing; a higher PV value indicates more dramatic speed fluctuations.

1.3 Latest Scientific Finding: 2023 Boston Marathon Big Data Analysis

In 2023, a study published in the European Journal of Sport Science analyzed over 100,000 finish records from the Boston Marathon between 2015 and 2019, dividing runners into five groups based on finish time (Sub-2.5, 2.5~3.0, 3.0~3.5, 3.5~4.0, 4.0+). The results showed:

This data reveals a harsh truth: the more dramatic the pacing fluctuations, the more significant the second-half speed decay, and the lower the lactate threshold utilization. In other words, stable pacing is not merely a display of “mental discipline,” but a necessary condition for the efficient operation of the physiological energy systems.

2. Core Mechanisms of Exercise Physiology and Biomechanics

2.1 Dynamic Interaction of Energy Systems: The Tug-of-War Between Aerobic and Anaerobic

At marathon intensity (approximately 75%~85% of VO₂max), the human body primarily relies on aerobic metabolism for energy. However, when pace surges momentarily (e.g., accelerating from 4:30/km to 4:00/km), the muscles’ demand for adenosine triphosphate (ATP) immediately exceeds the aerobic system’s supply ceiling, forcing the activation of anaerobic glycolysis.

Anaerobic glycolysis produces ATP at only 1/18th the efficiency of the aerobic system (each glucose molecule yields a net of 2 ATP via the anaerobic pathway versus 36~38 ATP via the aerobic pathway), and it is accompanied by the rapid accumulation of hydrogen ions (H⁺) and lactate. As intramuscular H⁺ concentration rises, it directly inhibits the activity of phosphofructokinase (PFK-1), creating a negative feedback loop that slows glycolysis. This forces the muscles to shift towards reliance on free fatty acid oxidation—but the ATP production rate from fatty acid oxidation is far lower than glycolysis and cannot meet high-intensity demands. Consequently, the runner experiences a debilitating “legs filled with lead” sensation.

2.2 The Exponential Amplification Effect of Pacing Fluctuations on Glycogen Depletion Rate

The critical point is: the glycogen depletion rate is not linearly related to speed, but exponentially related. According to the classic study by Hermansen and Stensvold (1972), when exercise intensity increases from 60% VO₂max to 85% VO₂max, the muscle glycogen depletion rate rises from 0.35 mmol per minute per kilogram of muscle to 1.2 mmol—a 3.4-fold increase. If intensity is further raised to 105% VO₂max (above the anaerobic threshold), the depletion rate can reach 2.5 mmol per minute per kilogram of muscle, a more than 7-fold increase compared to 60% intensity.

We can construct a simplified numerical model to illustrate the cost of pacing fluctuations. Consider a Sub-3 target runner (average pace 4:16/km, i.e., 14.06 km/h). If they complete the race with perfectly even splits, total glycogen consumption would be approximately (E_{even} = \bar{P} \times T), where (\bar{P}) is the average power output (approximately 250W) and (T) is the total time (approximately 2 hours 58 minutes). If instead they adopt a “fast-start, slow-finish” strategy—running the first 10km at 4:00/km (15.0 km/h) and the remaining 32.195km at 4:24/km (13.64 km/h)—the power output in the first segment would be approximately 270W, dropping to 240W in the latter segment. Since the glycogen depletion rate is exponentially related to power (( \dot{G} = a \cdot e^{bP} ), where (a) and (b) are individual constants), glycogen consumption in the first 10km would be 1.35 times that of even pacing. Although intensity decreases in the latter segment, the depletion rate only drops to 0.85 times that of even pacing. In total, the variable pacing strategy’s overall glycogen consumption is approximately 1.12 times that of even pacing—meaning at the 30km mark, the variable pacer may have exhausted their glycogen stores, while the even pacer still retains about 10% in reserve.

2.3 Disruption of Lactate Dynamic Equilibrium and the “Lactate Shadow”

Under steady pacing, lactate production and clearance rates reach a dynamic equilibrium (steady state), with blood lactate concentration maintained in a stable range of 2~4 mmol/L. However, when pace fluctuates dramatically, each acceleration causes lactate production to momentarily exceed clearance capacity, causing blood lactate concentration to rise in a “stepwise” fashion. Each acceleration is like pouring another cup of water into an already nearly full cup—even if you subsequently slow down, the lactate clearance rate cannot immediately recover, because the clearance mechanisms (primarily dependent on the liver and slow-twitch muscle fibers) require several minutes to reach a new equilibrium.

Specifically, suppose a runner accelerates from 4:30/km to 4:00/km for 1 kilometer at the 10km mark. Blood lactate concentration might spike from 2.5 mmol/L to 5.5 mmol/L. Even if they then decelerate back to 4:30/km, blood lactate would still require approximately 8~12 minutes to return below 3.0 mmol/L. For the remainder of the race, this excess lactate burden continuously suppresses fat mobilization (because lactate inhibits hormone-sensitive lipase in fat cells), forcing the body to rely more heavily on remaining glycogen—creating a vicious cycle.

2.4 The Biomechanical Cost: Speed Fluctuations and Muscle Work Efficiency

From a biomechanical perspective, running’s mechanical efficiency has an optimal range at specific speeds. Research shows that running economy follows a U-shaped curve with respect to speed: excessively slow speeds lead to shortened stride length and prolonged ground contact time, while excessively fast speeds increase vertical oscillation and braking impulse. When pace fluctuates dramatically, the runner is forced to frequently switch between stride frequency and stride length combinations, increasing the number of transitions between eccentric contractions (braking) and concentric contractions (propulsion). Each transition is accompanied by additional mechanical energy loss.

From the force-velocity profile perspective, the power output of the quadriceps and triceps surae peaks at specific angular velocities. When pacing fluctuations exceed ±5%, the shortening velocity of muscle fibers deviates from the optimal range, causing muscle fiber recruitment patterns to shift from efficient Type I (slow-twitch) fibers to inefficient Type IIa (fast-twitch) fibers, further accelerating glycogen depletion.

3. Key Parameter Measurements and Comparative Analysis

3.1 Even Splits vs. Negative Split vs. Positive Split

To quantify the physiological costs of different pacing strategies, we use a runner targeting Sub-3 (2 hours 58 minutes) as an example, simulating a comparison of three strategies. Assume their VO₂max is 58 ml/kg/min, lactate threshold pace (LT pace) is 4:05/km, and running economy is 210 ml/kg/km.

Table 1: Simulated Comparison of Physiological Parameters for Three Pacing Strategies

Table 2: Pacing Strategy Comparison for Sub-2.5 (2 hours 28 minutes) Elite Runners

3.2 Data Interpretation: The “Slight Advantage” of Negative Splits and the “Catastrophic Cost” of Positive Splits

From Table 1, it is clear that the difference in total glycogen consumption between even splits and a slight negative split is only 15 grams (approximately 3.7%). However, the positive split strategy’s total consumption reaches 485 grams—75 grams more than even splits. This is equivalent to expending approximately 1.5 hours’ worth of glycogen reserves. More critically, the positive split strategy leaves only 35 grams of glycogen at the 30km mark, meaning the final 12km relies almost entirely on fat oxidation. The maximum power output from fat oxidation can only support approximately 70% of VO₂max intensity—far insufficient to maintain a 4:27/km pace. This explains why positive split runners often encounter severe “hitting the wall” after the 35km mark.

The slight negative split (second half approximately 3%~5% faster than the first half) is recommended by many top coaches because it leverages the physiological characteristic of “increased fat oxidation ratio” in the latter half of a marathon. A slightly conservative pace in the first half (approximately 95% of lactate threshold speed) preserves glycogen reserves while allowing the body to gradually increase fat mobilization efficiency. Entering the second half, when fat oxidation has become the primary energy source, a moderate acceleration (to 98%~100% of lactate threshold speed) does not excessively deplete glycogen reserves, but instead utilizes the “physiological momentum” accumulated in the first half to elevate the overall average speed.

4. Periodized Training Plans and Practical Race Execution Guide

4.1 Training Philosophy for Building “Pacing Stability”

To achieve even splits or a negative split, runners must repeatedly practice calibrating “perceived pace” against “actual pace” in training. This requires substantial tempo runs and marathon pace (MP) runs to allow the neuromuscular system and energy metabolism systems to memorize the stable state at the target pace.

4.2 12-Week Sub-3 Breakthrough Plan (Weekly Basis)

The following plan is suitable for runners with a half marathon time within 1 hour 25 minutes and at least one full marathon finish experience. Intensity zones are defined using both heart rate and pace:

Key Plan Points:

• Tuesday Intervals: The goal is to improve VO₂max and lactate buffering capacity, but recovery time must be strictly controlled to ensure each repetition’s pace remains stable—no fast starts with slow finishes.
• Thursday Tempo Run: This is the most critical “pacing calibration day.” Wear a GPS watch and check pace error every kilometer, aiming to keep each kilometer’s error within ±3 seconds.
• Saturday Long Run: The “progression” in the latter part of the long run is key to simulating a negative split. Starting from the 20km mark, gradually accelerate to marathon pace (MP), maintaining MP for the final 5km without exceeding it.
• Recovery Run: Heart rate must be controlled in Zone 1 (55%~65% of maximum heart rate). The purpose is to promote blood circulation and metabolic waste clearance, not to provide training stimulus.

4.3 Pre-Race Taper Adjustments

Begin tapering three weeks before the race, reducing total mileage to 70%, 50%, and 30% of the original volume. During the taper, maintain Tuesday interval sessions (modified to 4×800m @ 3:50/km) and Thursday tempo runs (modified to 20min @ 4:25/km) to preserve neuromuscular “speed memory” while reducing glycogen depletion and muscle micro-damage.

5. Race Nutrition, Environmental Adaptation, and Race Execution Strategies

5.1 Quantified Strategies for Carbohydrate Loading and In-Race Nutrition

Three days before the race, perform the classic “glycogen loading” protocol: 3 days of low carbohydrate (5g/kg body weight/day), followed by 3 days of high carbohydrate (10g/kg body weight/day). For a 70kg runner, daily carbohydrate intake must reach 700 grams during the final 3 days. Two hours before the race, consume 1.5g/kg body weight (approximately 105 grams) of low glycemic index carbohydrates (such as oatmeal or whole wheat toast).

The golden rule for in-race nutrition: consume 60~90 grams of carbohydrates per hour (combining 6%~8% concentration sports drinks with energy gels). For a Sub-3 runner, it is recommended to refuel every 5km: 1 energy gel + water at 15km, 1 energy gel + electrolyte drink at 25km, and 1 caffeinated energy gel at 35km (caffeine can increase fat oxidation rate by approximately 20%~30%).

5.2 Environmental Adaptation: Pace Adjustments for Temperature and Humidity

According to World Athletics’ temperature correction model, for every 5°C increase above 15°C, pace should be slowed by 2%~3%. If race day temperature reaches 25°C, the Sub-3 target should be adjusted to 3:05~3:08. When humidity exceeds 70%, sweat evaporation efficiency decreases and core body temperature rises faster. It is recommended to increase hydration frequency to every 2.5km, with each hydration stop including additional sodium electrolytes (500mg of sodium per liter of water).

5.3 Classic Course Race Execution Strategies

• Dongjin Wuling (Elevation 0→3,275m): This is an ultra-long climbing race where the pacing variability coefficient will inevitably be high. However, you should aim to maintain stable “power output” rather than stable “pace.” It is recommended to use a power meter, maintaining power at 85%~90% of Functional Threshold Power (FTP) during climbs, avoiding excessive output on steep sections.
• One-Day Taipei to Kaohsiung / Twin Towers (Flat + Headwind): On flat sections, even pacing should be the highest guiding principle. Reduce power output by 10% in headwind sections, and return to target power in tailwind sections—never exceed target speed just because of a tailwind.
• Yangmingshan Wind Sword (Rolling Terrain): Use gravity to accelerate on downhills, but control cadence (maintain above 180 steps per minute). On uphills, maintain stable output based on “perceived effort,” avoiding dramatic heart rate fluctuations caused by gradient changes.

6. Common Operational Mistakes and Scientific Myth Busting

6.1 Myth 1: “Running the First Half Faster Buys You Buffer Time”

This is the most fatal suicidal strategy in marathon running. Running 10 seconds/km faster than target pace for 10km means consuming approximately 20 grams of glycogen in advance—enough to support 3km at target pace later in the race. When you inevitably slow down in the second half due to glycogen depletion, the time lost is often 2~3 times the time “gained” in the first half. Scientific data shows that positive split runners experience an average speed decay after the 35km mark that is 4.2 times greater than negative split runners.

6.2 Myth 2: “Lactate is the Culprit of Fatigue; Avoid Lactate Accumulation”

Lactate is not a metabolic waste product but an important energy carrier. Through the “lactate shuttle” mechanism, lactate can be transported to the heart, slow-twitch muscle fibers, and the brain as fuel. What actually causes fatigue is muscle acidification from hydrogen ion (H⁺) accumulation, not lactate itself. Therefore, rather than fearing lactate, you should train to improve your body’s “lactate clearance rate” and “buffering capacity”—this is precisely the core purpose of interval training and tempo runs.

6.3 Myth 3: “Accelerating in the Final 10km Can Create Miracles”

This “killer blow” strategy is only suitable for top elites (Sub-2.2), because their aerobic reserves are extremely high, allowing them to tap into anaerobic reserves in the final stages without collapsing. For Sub-3 or Sub-3.5 runners, acceleration in the final 10km is often constrained by glycogen depletion. A more scientific strategy is to execute a “micro-surge” at the 30km mark (5 seconds/km faster than target pace, sustained for 2km) to stimulate the nervous and cardiorespiratory systems, then return to target pace. This is far more effective than an all-or-nothing gamble in the final stages.

6.4 Myth 4: “The Lower the Pacing Variability Coefficient, the Better—Aim for Zero Fluctuation Throughout”

While a low PV value is beneficial for energy efficiency, overly rigid pacing (PV<1%) also ignores real-time changes in terrain, wind direction, and body condition. In actual races, slight fluctuations (PV between 2%~4%) are normal. The key lies in the “symmetry of fluctuations”—the magnitude of accelerations and decelerations should be roughly equal, and the average pace must fall within the target range. True masters are not those with “zero fluctuation throughout,” but those who “fluctuate yet always remain within the target zone.”

7. Expert FAQ

Q1: How do I calculate my own Pacing Variability (PV)? What tools do I need?

You only need a GPS-enabled sports watch with automatic per-kilometer splits. After the race, export the per-kilometer pace data to Excel or TrainingPeaks. First, calculate the average pace for each 5km segment (8 segments plus the final 2.195km), then apply the formula (PV = (Standard Deviation ÷ Average Pace) × 100%). If your PV value exceeds 5%, your pacing stability is insufficient and you should strengthen tempo run training. If it exceeds 8%, you need to re-examine your race strategy and consider whether you started too fast.

Q2: How “negative” should a Negative Split be? Is there an ideal range?

Research shows that for amateur runners, a second half that is 1%~3% faster than the first half is the ideal range. A negative split exceeding 5% typically indicates an overly conservative first half that fails to fully utilize energy reserves, which can actually lower overall performance. For Sub-3 target runners, it is recommended to complete the first half at target pace +5 seconds/km (i.e., 4:21/km) and the second half at target pace (4:16/km), resulting in an overall negative split of approximately 2%. Please note that a negative split is not about “deliberately slowing down,” but rather “restraint in the first half, release in the second half.”

Q3: On courses with significant elevation changes (such as Yangmingshan Wind Sword or Dongjin Wuling), the PV value will inevitably be high. How should I adjust my strategy?

On undulating courses, you should shift your monitoring from “pace” to “power.” A running power meter (such as Stryd) can measure your power output in real time. It is recommended to maintain power at 85%~90% of Functional Threshold Power (FTP). On uphills, allow pace to naturally decrease, but power must not exceed the limit; on downhills, use gravity to accelerate, but power can drop to 75%~80% of FTP to reduce muscle strain. This strategy of “stable power, fluctuating pace” effectively reduces glycogen depletion rates.

Q4: What is the difference between even splits and negative splits for breaking Sub-2.5?

For Sub-2.5 elite runners (average pace 3:33/km), aerobic system utilization already exceeds 90%. At this level, even minor pacing fluctuations directly impact lactate dynamic equilibrium. Most top athletes (such as Eliud Kipchoge) adopt a strategy of “even splits + acceleration in the final 3km” rather than a full negative split. The reason is that at extremely high intensities, being “conservative” in the first half means losing the advantage of staying with the lead pack, and the body’s ability to adapt to high-lactate environments is stronger. Therefore, Sub-2.5 runners should primarily use even splits, only allowing pace to increase to within 3:25/km in the final 5km.

Q5: Does in-race carbohydrate intake affect pacing stability?

Absolutely. Dramatic fluctuations in blood glucose levels (alternating hyperglycemia and hypoglycemia) directly affect the central nervous system’s “perceived exertion” and the muscles’ fuel supply. It is recommended not to wait until you feel thirsty or hungry to refuel, but to consume small amounts of liquid carbohydrates at fixed intervals (every 15 minutes, totaling 60~80 grams per hour). Additionally, avoid consuming more than 50 grams of carbohydrates at once, as this can cause gastrointestinal discomfort and excessive insulin secretion, leading to a “sugar crash” phenomenon where blood glucose rises then falls sharply—this severely disrupts pacing stability. The ideal refueling rhythm is: alternate between energy gels and electrolyte drinks every 5km, ensuring blood glucose remains in a stable range of 4.5~6.0 mmol/L.

---

Conclusion: The Pacing Variability Coefficient is not just a number on a data dashboard; it reflects the “discipline” of an athlete’s energy metabolism system. On the journey towards Sub-3 or Sub-2.5, every pacing fluctuation invisibly depletes your most precious glycogen reserves. Remember: the marathon is the ultimate test of “patience” and “science.” Stable pacing is not conservatism—it is the deepest respect for your body’s energy systems. When you can control your PV value within 3%, you have mastered the most powerful weapon for breaking your personal records.