Competitive swimming has been part of the Olympic movement ever since the first modern games in 1896. Currently, it is one of the largest sports in the Olympics with 32 events (10.5% of all Rio 2016 events) being contested between both sexes. Similar to many other time-trial sports, such as track, cycling and rowing, the objective is to cover a set distance in the least amount of time and before your competitors. These events range from the 50 y sprint (American collegiate level) to the 1,500 m freestyle in the pool (Olympic level), while the 10,000 m freestyle is contested in open water swimming. The breaststroke, backstroke, and butterfly events are also competed in the pool and are limited to the 50 (only at World Championships) through to the 200 m distances, while the individual medley races are limited to the 200 and 400 m distances (100 m only at short course meets). Event durations range from about 17 s (50 y males) to about 16 minutes in the 1650 y mile race at the collegiate level. Over 52% of all events last around 5 minutes or less including all strokes.
Back when I was an assistant swimming coach in my days at the Pontifical Catholic University of Puerto Rico, I was introduced to the energy zone systems by my Head Coach, Ralph Ramirez. These were categorizations of different intensities from which we expected different physiological adaptations to while distributing the total training volume within them. Energy zone systems were initially suggested based-off from lactate curves during an incremental intermittent swimming test (Figure 1). In the 2012 Journal of Swimming Research article written by Hall of Fame Swimming Coach, Ernie Maglischo, an attempt was made to further detail on this by describing how force and power aspects but not speed lead to greater muscle recruitment (Figure 2), and hence production of lactate, especially fast twitch muscle fibers being recruited at higher intensities which led to the designated six distinct training intensity zones (Maglishco, 2012).
Figure 1. Training velocity zones suggested from the lactate curve of a multi-stage incremental swimming test (taken from Maglishco, 2012).
Figure 2. Revised training velocity zones suggested from the fiber-type muscle recruitment and heart rate (taken from Maglishco, 2012).
Measuring the maximal oxygen consumption (VO2max) is considered the gold standard measurement of aerobic capacity. For many years, practitioners prescribed training based-off percentages of the VO2max. However, the nature of the sport of swimming makes this a complicated task due to the tedious protocols, expensiveness, and very time consuming process. Add to that the invasiveness and lack of expertise of lactate threshold (LT) measurements which are typically rejected within collegiate swimming coaches on a wide scale basis and rightfully so due to budgets and collegiate training time restrictions. Hence, coaches tend to gravitate towards more traditional "testing set" methods.
One such method, is the "Best Average Swimming Test" (Bavg). The Bavg was introduced by legendary Hall of Fame Swimming Coach, Jon Urbanchek. It was originally performed using a set of 5x400 m swims with only 1-minute of rest in between repetitions while maintaining your best average pace, hence the name (Figure 3). The resulting average of the speed throughout the test was used as a threshold (Bavg) which would later be used to devise training zones in what Coach Urbanchek called the "Color System", given by the color of a swimmer's face when swimming sets at different intensities. Later on, other coaches manipulated the distance repetitions and total volume of the test. This inevitably changes the metabolic response that an athlete experiences while performance the test. Moreover, athletes with different physiological performance profiles may be hindered or benefitted depending on which particular test structure is used (Sandford & Stellingwerff, 2019).
Figure 3. Different set structures of the Bavg performed by coaches.
Looking back, despite the understanding of the training zones by coaches, there was still missed context of those different intensities or zones from an individual manner and the true demands of the event. Some great work developed by Dr. Gareth Sandford analyzing the physiological performance profiles of a homogenous group of elite 800 m middle-distance runners illustrated the misunderstood nature of the traditional research done within this particular group (Standford & Stellingwerff, 2019; Sandford et al., 2019). By specifically pointing out the absence of reporting on the maximal sprinting speed (MSS) in classic aerobic training studies, this created an unconscious bias that inherently missed out on important context when analyzing athletic performance profiles by not accounting for the anaerobic speed reserve (ASR). The ASR is defined as the difference between MSS and an athlete's MAS or velocity associated to the VO2max (Sandford et al., 2021). Hence, this prior misunderstanding of context could mean that when classic research study designs prescribed aerobic training based-off a percentage above the MAS with no consideration of the athlete's ASR, there could possibly be a misrepresentation of expected physiological responses created by unwanted variability. When this happens, we are not optimizing training to pinpoint appropriate adaptations along with erroneous analysis.
One of the most common testing set structures to obtain a Bavg speed, especially in collegiate swimming has been the 10x100 m (Figure 3). The inter-set rest interval is dependent on the send-out time, and since each athlete will arrive at different times, this will be variable. The ability to understand at which percentage of the ASR the athlete is swimming at during the Bavg would be obscured by the variable inter-set rest intervals for each athlete, which by themselves are obscured by the athlete's ability to recover within each of those time frames. Also, to measure the MSS we would need to perform additional sprint testing (i.e., 25 m for time), taking up more time than the already lengthy Bavg protocol (15 minutes each athlete). A team with a total of 32 athletes and an 8 lane swimming pool would take up to 1 hour and 40 minutes to complete for top level athletes at best. Then, add the sprint testing and this becomes restrictive in collegiate swimming.
However, recently a three-minute all-out swimming test (3MT) was been validated and shown to be reliable to measure the critical speed (CS) in swimming (Thomas & Tsai; Piatrikova et al., 2019). The CS is a mechanical measure of maximal aerobic steady state demarcating heavy and severe intensities (Pettit et al., 2018). It demarcates sustainable and unsustainable intensities of exercise, which means that intensities below it can be sustained indefinitely (theoretically) and intensities above it are finite while at CS intensity can theoretically be maintained around 30-60 minutes.
The second metric we can obtain from this test is the distance capacity available above the CS (D'; pronounced D-prime). Together, these two metrics can be used to predict the tolerance of intensities above the CS (Pettit et al., 2012). This makes it possible to predict times of events that require intensities above the CS (Pettit et al., 2018) and interval work times for swimming sets (Piatrikova et al., 2020). The theoretical background of the test stems from the seminal work by A.V. Hill in the 1920's, when he first demonstrated a plot of the hyperbolic relationship of the distance-velocity curve of world record runs showing what appears to be an asymptote for which the same can be observed in swimming (Figure 4). Since the 3MT requires an all-out maximal effort from the beginning, we could also measure the MSS within the first lap split time (at feet touch), hence, we could also calculate the ASR by using the CS as the MAS measure (Figure 5).
Figure 4. World records as of June 2021.
Figure 5. Theoretical representation of the 3MT time-velocity profile.
Originally, the 3MT was analyzed by measuring the average speed of 15 s intervals with the last 30 s being used to calculate the CS. However, in my dissertation work, I found that using a simple stopwatch with split times we could capture the same speed drop-off phenomena as in the time-interval analysis (Figure 6) independent of athlete performance level (Figure 7) with high inter- and intra-rater reliability. In the time-split method, averaging the last two laps (25 m pool) was not significantly different from the original CS analyzed by time-intervals. However, despite the differences in the D', the lap split-method showed higher values than than the time-interval method, which in a conversation with Dr. Piatrikova, who previously had validated the 3MT in swimming, had mentioned that this value assimilated the values of conventional multiple time trial estimations of the CS (i.e., plotting of the time-speed relationship of multiple trials from 100 - 1500 m swims).
Figure 6. Mean speed profile of the 3MT analyzed as time-interval (solid line) and lap-split method (dotted line) for one athlete.
Figure 7. Mean speed profile of the 3MT analyzed as lap-split method for four athletes of differing performance levels.
What should be more enticing for swimming coaches is that we could also build training zones based of off individualized profiles in which each zone can target a physiological adaptation. Burnley and Jones (2007), categorized training zones based-off of physiological landmarks that demarcate unique amalgamations of physiological responses (Figure 8). Futhermore, Piatrikova and colleagues showed that the 3MT could be used as an alternative to the traditional testing procedures to estimate the LT where resources and time might be limited (Piatrikova et al., 2018). By multiplying the CS (shown as CP in Figure 8) by 0.90 we could have the lower boundary of the heavy intensity zone (LT; shown as GEP in Figure 8). The LT becomes the upper boundary for moderate intensity exercise. The extreme zone will then be demarcated by the intensity of which fatigue is reached prior to attaining VO2max, these are intensities that typically can be sustained for less than 120 s.
Figure 8. Taken from Poole & Jones 2012.
Although more research should be done to identify the lower boundary of the extreme zone in swimming, Pettit found that the VO2max was first evoked at the 90 s mark of the 3MT in running (Pettit et al., 2012). Within my conversation with Piatrikova, she mentioned that this could also be the average speed of the first 200 m during the swimming 3MT. In fact, in my dissertation work, I used a criteria to exclude athletes that did not give maximal effort in the 3MT. The criteria consisted of excluding those swimmers who's CS were greater than the mean speed of the 100 & 125 m splits. The justification for this criteria relied on the fact that the athletes averaged between 68 and 78 s for those distances and the transition for the body to predominantly use aerobic energy contribution is considered to be around 75 s. Due to the increasing metabolite accumulation from maximal effort exercise, an athlete should not be able to accelerate at faster paces beyond this time point during the 3MT. So in my opinion, during post-analysis we could choose the average speed of a distance that is closest in time to the 75 s mark when performing a 3MT with your own athletes.
For a detailed explanation on the distinct physiological environments in each zone for healthy individuals, you are best referred to Burnley & Jones (2007) and Poole & Jones (2007) reviews listed in the references. Briefly, in the moderate zone, a steady state of VO2 can be achieved within 3 minutes with fatigue mainly resulting from hyperthermia, reduced central drive or motivation, and/or muscle damage. In the heavy zone, a VO2 slow component is observed with a delayed steady state of VO2 within 10 - 20 minutes with fatigue mainly resulting from glycogen depletion and hyperthermia. In the severe zone, VO2 slow component is evident and develops continuously if intensity is below VO2max with no steady state until VO2max is reached. Fatigue in this zone is due to the depletion of finite energy stores that represent the D' and accumulated metabolites leading to muscle contraction failure. Within the extreme zone, there is no VO2 slow component and accumulated metabolites leading to muscle contraction failure prior to reaching VO2max (Figure 7).
Coaches may ask, is CS better than the traditional Bavg? In my dissertation work, I found that there is high variability of ASR profiles within individual athletes of a swimming team. Figure 9 further shows the variability of the percentage at which the Bavg pace is represented within each individual ranging from -2% to 57% of an athlete's ASR. Athlete Q was highlighted in red since this athlete was the only one with a negative percentage value. This athlete also happens to have the lowest MSS and smallest ASR meaning that a complete depletion of the D' occurred both during the Bavg and the 3MT. If we try calculating the ASR using the Bavg pace instead of the CS, we may actually be training harder than we would anticipate. This could risk targeting other physiological adaptations that are not intended by shifting into higher training zones than those desired. Figure 10 illustrates this by looking at an example athlete with the Bavg pace representing 29.5% of the true ASR (calculated by CS). In such a scenario where the athlete trains at Bavg pace, this would not be as sustainable as when using the CS method.
Figure 9. Individual variability of ASR and percentage of ASR at which BAST is expressed at.
Figure 10. Comparison of an example athlete's ASR when using both thresholds (BAST & CS) as the lower boundary.
Dr. Eva Piatrikova used the 3MT to set individualized training sets throughout a 15-week season for highly trained swimmers who ranged from 78 - 90% of the world record in their main events (Piatrikova et al., 2020). She found that swimmers increased CS alongside a reduction of the D' with trivial changes to MSS. Swimmers also experienced improvements of 1.2 and 1.6% of their main two events, respectively. These improvements were attained despite a reduction of 25% in training volume. A similar pattern for the CS, D' and MSS was observed during the 3MT assessments in my dissertation during a 13-week season (similar % of WR level athletes). Although in our study we did not quantify the volume performed for the prior year, the average of volume was 20.6 km per week with a range of 4.4 km in the first week and 32.9 for the highest week (set by the Head Swimming Coach).
If we further want to use the results of a 3MT for profiling analysis, we could use the ASR to distinguish between a homogenous group of swimmers similar to what Dr. Sandford did with elite middle-distance athletes (Figure 11). He noticed the difference between mechanical metrics (MSS and MAS - shown as vVO2max in Figure 11) despite all athletes having similar VO2max while all having best times and within 1.3 s of each other. In swimming, we could either use the CS or MAS as the lower boundary of the ASR. Given that swimmers are familiarized with maximal efforts, a 3MT would make sense especially considering that the 3MT has already been shown to be a valid and reliable performance test. From the 3MT we can calculate other physiological boundaries, such the MAS as explained above.
Another way to calculate the ASR is by swimming an all-out 400m swim. This has a time and effort approximation to Dr. Sandford's recommendations of using the 1500 m run or a 5 - 6 minute maximal effort time trial. There has been abundant research in swimming which have calculated the CS by way of linear regression models from multiple time-trials of differing durations (Wakayoshi et al., 1992) and a simple percentage of the average speed of the 400 m all-out (Arroyo-Toledo, 2018). The latter would be best given that it is a single short test (~ 4 - 6 depending on athlete level) resulting in the MAS and you could choose to calculate the CS as 80% of that speed. Which metric to use could be up for debate, but a couple of important facts are that the phosphocreatine system (ATP-PC) only regenerates below the CS through aerobic metabolic pathways and the MAS (vVO2max sustainable for around 4 - 6 minutes) is a mechanical intensity that sits within the severe intensity zone (shown in Figure 8) below the lower boundary of the extreme zone (fatigue before reaching VO2max).
In my opinion, I believe a 3MT precisely measures the CS - which is important at precisely identifying sustainable and unsustainable intensities. Then, we could choose to calculate MAS from the resulting CS of the 3MT as performed by Piatrikova (2020). Taking the information from the running 3MT performed by Pettit (2012) and what I found in my dissertation, the average speed of the first 75 - 90 s could be used as a surrogate for MAS. Nonetheless, more research is needed to verify the phenomenology acting behind these performance intensities. One question that does arise from this chain of thought is, could the first 75 - 90 s average speed of the 3MT have any association to the average speed of the 400 m? These are questions that should be investigated with further research.
Figure 11. ASR profiles of elite middle-distance female runners taken and modified from Sandford & Stellingwerff (2019). Description from original graph: Anaerobic speed reserve and velocity at 4 mmol/l lactate (v@4 mmol/l) across three elite middle-distance female profiles from each of the 800m sub-groups tested in 2017. Note the between individual diversity across v@4 mmol/l, vVO2max and Maximal sprint speed—despite all having a season’s best over 800m within 1.3 s of each other.
Moreover, we could also use these profiles to set racing pace strategies (Maugher et al., 2014). An athlete with a high CS but low MSS (small ASR), may be able to start out a race faster than an athlete with a lower CS but higher MSS (big ASR; negative pacing). Training sets may also differ between these two athletes. One proposed method to prescribe high intensity interval training (HIIT) sets was used by Piatrikova in Equation 1. Basically, whenever you train within the ASR you are depleting the D'. The rate at which you deplete it is dependent on the intensity. So, the higher the percentage of D' you use the faster the split will be, the more taxing the interval becomes, and the more you drain the D'. Athletes with high CS and low MSS may be best served with long interval HIIT sets (60 - 240 s) while those with low CS and High MSS may enjoy more short interval HIIT sets (15 - 60 s). To go deeper into how to use this for training purposes should be left for another article, however, the reader is referred to an excellent source in the book by Luarsen & Buchheit titled "Science and Application of High Intensity Interval Training: Solutions to the Programming Puzzle" or dive into some of Dr. Sandford's work (referenced below).
Equation 1. Interval time = [distance - (D' x % of D')] / CS
So how do we administer a 3MT? It is very simple, but it depends on your pool dimensions and your resources. First thing to consider is the obvious, if the 3MT is performed in a short course pool the results are confined to those pool dimensions as it could over estimate long course CS training paces, since the added wall pushes increase opportunities for acceleration (and vice-versa). Researchers have used stopwatches to measure splits at feet contact, similar to how coaches already take lap split times. The last two completed laps in a short course and the last single lap of a long course 3MT should suffice to measure a true CS. If you have additional staff to take splits for all athletes then that is great. If you are the only person on the deck, don't worry. For my dissertation and other research studies, a video recorder was used to analyze retrospectively these split times. I also found that this method had excellent inter-rater reliability, which means that the measurements were nearly the same within two researchers and its a reliable method. One swimmer should be allowed in a lane at a time and as many swimming lanes (i.e., swimmers) can be used to perform the 3MT at a time as long as a complete view of the pool or lanes are captured on film.
Swimmers themselves could set up a camera on one side and perform a do-it-yourself 3MT. Athletes should absolutely not be aware of the remaining time to avoid any pacing throughout the 3MT. They should begin at all-out pace from the beginning on the command of "GO", a whistle or however you'd like. The test ends after the 3 minutes have concluded on a whistle, scream or touch to stop the athlete. Just make sure you have someone in charge and aware of the time to stop you on time when the required time elapses because this not for the faint of heart. For example, the athlete in the Video 1 might have or might have not gone for an additional grueling 30 s. Hopefully right after the 3 minutes and not after someone remembers during a conversation outside the pool after the video stopped recording.
Video 1. A swimmer performing the 3MT in a local gym club lap pool.
Finally, one must consider when to do such a test during your programming and periodization models. The 3MT is a very fatiguing and neuromuscular taxing test. So, don't try it on a rest day! In my dissertation, I found that you can see changes in as little as 3 weeks, but performing it every 6 weeks is fine. Just make sure that every time you perform the test you have performed a standardized warm-up and it is within a similar context of the training week from the last time it was performed. This means that if the first time you did it on a Monday you may not want to repeat it on a Saturday, since these training days may not be the same in your training schedule (i.e., Sunday may be "off day" while Thursday not).
In conclusion, the 3MT can be a more accurate and time-efficient performance test to measure the CS which can be used to set your individualized training zones. This should be more appropriate than the general and traditional benchmarks that have been used historically in the past and are still used today in swimming. In future articles, I will try diving into how to prepare training by using these zones.
References
1. Arroyo-Toledo, JJ., Huusgaard JÁ, Joensen, PV, Hofgaard, DM, & Jákupsstovu, RI (2020). Prescribing Aerobic pace training to the National Swim Team of Faroe Islands. International Journal of Scientific & Engineering Research. 9 (4): 1228-1231.
2. Hellard, PP (2018). Dynamics of the metabolic response during a competitive 100-m freestyle in elite male swimmers. International Journal of Sports Physiology and Performance, 13(10): 1011-1020.
3. Luarsen, P & Bucheit, M (2019). The Science and Applications of High Intensity Interval Training: Solutions to the Programming Puzzle, Human Kinetics, Champaign, Ill.
4. Maglischo, E. (2012). Training Zones Revisited. Journal of Swimming Research, 19:2.
5. Mauger, AR, Neuloh, J & Castle, PC. (2012). Analysis of pacing strategy selection in elite 400-m freestyle swimming. Medicine and Science in Sports and Exercise, 44(11): 2205-2212.
6. Pettitt, R., Jamnick, N., & Clark, E. (2012). 3-min All-out Exercise Test For Running. International Journal Of Sports Medicine, 33: 426-431.
7. Piatrikova, E., Sousa, A., Gonzalez, J., & Williams, S. (2019). Validity And Reliability Of The 3-minute All-out Test In National And International Competitive Swimmers. International Journal Of Sports Physiology And Performance, 13(9): 1190-1198.
8. Piatrikova, E., Sousa, A., Willsmer, N., Gonzalez, J., & Williams, S. (2018). Demarcating Exercise Intensity Domains In Freestyle Swimming: Is There And Alternative To Incremental Step Test And Beats Below Hrmax Method? . Journal Of Strength And Conditioning Research, 34(9): 2575-2584.
9. Piatrikova, E., Willsmer, N., Sousa, A., Gonzalez, J., & S, W. (2020). Individualizing Training In Swimming; Is There A Place For The Application Of The Critical Speed And Critical Stroke Rate Concepts? International Journal Of Sports Physiology And Performance, 15(5): 1-8.
10. Poole, DC. & Jones, AM. (2012). Oxygen Uptake Kinetics. Comprehensive Physiology, 2(2): 933-996.
10. Rodriguez, FA, & Mader, A (2011). Energy systems in swimming. Ch. 11, In book: World Book of Swimming: From Science to Performance. Ed. Seifert, L, Chollet, D & Mujika, I. Nova Science Publishers, Inc. Hauppauge, New York.
11. Sandford, G., & Stellingwerff, T. (2019). "Questioning your Categories": the Misunderstood Complexity of Middle-Distance Running Profiles with Implications for Research Methods and Application. Frontiers in sports and active living, 1(28): 1-8.
12. Sandford, g., Kilding, A., Ross, A., & Luarsen, P. (2018). Maximal Sprint Speed and the Anaerobic Speed Reserve Domain: The Untapped Tools that Differentiate the World’s Best 800 m Runners. Sports medicine, 49(6): 843-852.
13. Sandford, GN., Laursen, PB., & Buchheit, M. (2021). Anaerobic Speed/Power Reserve and Sport Performance: Scientific Basis, Current Applications and Future Directions. Sports Medicine, 51(10): 2017-2028.
14. Tsai, M., & Thomas, S. (2017). 3-min All-out In Swimming. . International Journal Of Sports Physiology And Performance, 12(1): 1-17.
15. Wakayoshi, K., Ikuta, K., Yoshida, T., Udo, M., Moritani, T., Mutoh, Y., & Miyashita, M. (1992). Determination and validity of critical velocity as an index of swimming performance in the competitive swimmer. European Journal of Physiology and Occupational Physiology, 64(2): 153-157.
Thanks for sharing your knowledge with others!