To increase transparency, PeerJ operates a system of 'optional signed reviews and history'. This takes two forms: (1) peer reviewers are encouraged, but not required, to provide their names (if they do so, then their profile page records the articles they have reviewed), and (2) authors are given the option of reproducing their entire peer review history alongside their published article (in which case the complete peer review process is provided, including revisions, rebuttal letters and editor decision letters).
Thank you for your careful efforts to deal with the second round of comments.
Please take into account the suggestions of reviewer 2, which should in principle be doable but will need some work, and have a look at the recommendation I had given in my first reply.
Clarity of the manuscript has increased following the revision. The main content has not been modulated extensively, nor are the conclusions in any way different. They continue to conform to the structure of the underlying data. The rewriting of the introduction has strengthened this section.
The authors have addressed all the points remonstrated in the first review and provided adaptations or replies. This has to a certain extent improved the manuscript. However, the lack of context for the reported observations increases the difficulty of interpretation and decreases the usefulness of this publication. Within this framework I remain of my former opinion that this type of data is important for the field, however it needs to be collected and reported in a conscientious manner with as much surrogate and corroborate data as possible, especially those data that are known to affect pacing strategy.
The authors have made most of the recommended corrections, other than having clarified some critical parts. However, a couple of points still require minor revisions.
Although substantial amendments have been made following the suggestions of the two Reviewers, this procedure has partially disrupted the flow of the article. Specifically, from the Paragraph 2 the article starts focusing on ultramarathon (UM) distances without a declared motivation (perhaps the increasing popularity of the events associated with the paucity of studies available on the topic?). Importantly, the aims of the study (Paragraph 4, lines 72-76) are not solidly supported by the previous paragraphs since the part about the effects of fatigue as a function of gradient category has been described only in the following phrase (Paragraph 4, lines 76-80). In this regard, I would suggest to put that part at the end of the Paragraph 3, before the enunciation of the study purposes.
Despite the article from Angus & Waterhouse (2011) (Paragraph 2, line 59 of the previous version of the manuscript) has been properly removed from the introduction, it is still present in the discussion (Paragraph 21, line 268-269 of the current version of the manuscript). Since in that article the authors did not observe major differences in pacing variation between the three cohorts of running speeds assessed on UM distance, I would recommend to remove it or replace it with Hofmann (2014).
This is an interesting article in an area that seems to be under-researched. The reviewers have done a very good job to make recommendations for improvement. I think that in future studies it would also be very interesting to study the evolution of step length and step frequency.
The manuscript could benefit from an editorial trim in some sections. Other than that it conforms to the guidelines.
The authors present an observational paper describing pacing strategies of ultra-mountain endurance athletes over a race distance of 173km (n=10) and 103km (n=5). Time and 3D-position data were recorded using GPS devices (0.2Hz) throughout the duration of the run. Data was distance-normalized (103km = 60%), filtered, centralized (group average speed) and divided into 5% bins. For each bin, relative uphill (>2.5%), downhill (<-2.5%) and level speed (-2.5<LEV<2.5%) were calculated. Over the race, speed decreased in all inclinations, albeit at different rates. A speed reserve was observed in the last bin for some inclinations. The level inclination demonstrated the highest speed variability and also the earliest switch from positive to even pacing. The most significant findings are described as a lack of negative correlation between predictors of performance (time stopped, speed loss, and speed variability) performance (time to finish).
The field of UM marathons is unique in offering a window on fatigue processes in grand magnitude race events. The acquisition of reliable continuous position and (psycho)physiological data is challenging and there is an enormous amount of confounders to be taken into account. The authors have done a labor intensive job collecting and cleaning GPS data from so many participants over such a long distance.
The contribution to the current base of knowledge lies mainly in the addition of a longer data set compared to a previous paper from the same group which assessed more outcome variables. The observations are subtly different between the two papers, begging the question if pacing is not more a function of the individual race profile and terrain than in any way indicative of a fatigue model. Also the sensitivity and specificity of pacing as a marker of fatigue may be severely compromised in these types of outdoor sports in which the environmental constraints (temperature, weather, trail condition, exposition, importance, etc.) may prove strong determinants of speed fluctuations. Placing the reported observations into the context of current models of fatigue and interpreting them within this framework can significantly strengthen the paper and make it more interesting. The authors are experts in this area having published multiple papers on similar topics.
The reported measures relate only to the dynamics of velocity management. There are no assessments of other fatigue indices on neither a physiological nor subjective level. This is disappointing as it would provide more relevance to the reported outcomes. I strongly agree with the authors’ final conclusion that “future studies are also warranted to investigate the importance of variables related to participant experience (number of years of practice, number of starts at a certain distance) in order to further characterise pacing and performance in UM events.”. More velocity data from different types of UM is needed to generate and validate a comprehensive framework for fatigue development in UM events and therefore this is a useful addition.
Why would you expect pacing to vary as a function of inclination? While it seems obvious that different inclines will result in different initial speeds, I believe a paragraph is warranted describing the framework that leads you to believe that speed dynamics evolve differentially in different inclinations.
How have you taken into account carry-over effects from earlier inclinations? Should this be mentioned in the limitations if not taken into account?
(Townshend, Andrew D. and Worringham, Charles J. and Stewart, Ian (2009) Spontaneous pacing during overground hill running. Medicine and Science in Sports and Exercise, 42(1). pp. 160-169.)
In the discussion once again I feel that there is not enough discussion (or even speculation) concerning the latent factors driving differential pacing on different inclines.
One of the novelties in the reported results lies in that all participants started under the same environmental conditions with the same aim, however some failed to complete that aim. More detail on the groups could make this into a much more interesting paper.
Do you believe that speed variability in the LEV condition is not driven by minor (indistinguishable) changes in gradient? As base velocity is higher, it might be expected that a given gradient change may have greater impact as compared to uphill for example.
Line 48/49: Stating that these ultra-long type of events enable the study of fatigue is I believe slightly too general. The fatigue that is studied in these events is specific to the nature of the event and also to the rather specialized population.
Line 62: Please make it clearer in the first sentence that these findings are from a different study.
Line 66-68: Are the populations comparable in terms of expertise, performance and motivation between these two studies? Are the course characteristics comparable (exposition, temperature, environmental constraints (altitude, snow, etc.)?
Line 76: The causality of this seems inverted: Please turn around to state that “the aim is to assess pacing, therefore we will use GPS” and not “because we have GPS, we will investigate pacing”.
Line 85: Maybe mention the race name so people can assess what kind of terrain it was conducted in?
Line 118: If you replaced all the missing values with 0, that would impact your means and your variability.
Line 121: “This procedure limited” might be more elegant.
Line 122-24: Maybe “zero-speed values associated with checkpoint locations were exempt from this treatment.”. I’m guessing you excluded these sections before smoothing?
Line 154-156: I find this very confusing. I’m guessing that it to some point reflects the “sparseness” of some datasets? In some cases it would be interesting to know how many points of each type are in each bin. It seems that the statistical analysis may be biased in those bins with a large skew toward any inclination.
Line 179: How about climate conditions? It would be good if you could give the temperature bounds and precipitation as temperature will impact pacing and precipitation will make the terrain more difficult to navigate.
Line 180: I think some demographics of the subjects should be presented – at minimum their level of experience with trail running and their age. Also whether they had prior knowledge of the course. I’m guessing they mere highly motivated, however is there any measure of their risk-taking behavior? This might be interesting for interpreting descent speed on technical terrain.
Line 181: I’m assuming all participants did aim to complete the 173km? Otherwise, seeing that we are taking into account anticipatory pacing and feed forward regulation this would be expected to have had an effect...
Line 202: In the study you are not just reporting, but also collecting data… Also line 203 please revise wording
Line 205: Overall decrease in speed
Line 207: “main component was greatest” – doubled expression?
Line 214: What are the further implications of the groups demonstrating no difference in pacing? Just a few ideas: If the 103km group would be designated as “lower performers” and the 173km group as “higher performers”, would this not indicate that pacing is completely insensitive to performance level? Why did the runners stop at 103km? Tactical decision? Were they fully fatigued? Were they equally fatigued as the 173km group? Did they pay a higher cardio-vascular price to adhere to the pacing profile? How about their absolute speeds – were these different? …
Line 219: No “a”
Line 227: In figure 2, there seems to be a significant speed increase in the last 10% of LEV and the last 5% of UH - Would this not constitute the speed reserve? Also, from the race profile in figure 1, the last DH segment looks steeper than the other ones. Do you not believe that the inclination and technicality of the last segment may have impacted the speed more than any pacing strategy?
Line 254: I don’t see how this paper provides a “basis for future studies of ultra-long duration exercise”. To my comprehension, the basis/framework has been provided in other publications and this is an addition of observational data.
This is a quite interesting and well-written article that expands the knowledge on some aspects of a still poorly studied research area. The introduction and background are adequate and figures and tables are complete and informative, although the large number of data represented in the figures makes their reading not always immediate. However, there are some points I would like to discuss with the Authors:
Paragraph 1, lines 34-36: Although the statement is correct, it should be noted how the principal reason of studying pacing strategy relates to the understanding of physiological and regulatory processes in function of the optimization of exercise performance, which does not only depend on bioenergetics. I would suggest to the Authors to rewrite this phrase, also in consideration that the article does not focus on the metabolic aspects of pacing.
Paragraph 1, line 40: I do not have access to the full article (Firth 1998).
Paragraph 1, line 42: Please note that in the reported study (Tucker et al. 2004) the Authors measured pacing strategies only on 800-, 5000- and 1000-meter distances. Moreover, I am not sure that defining the running dynamics observed in the 5000- and 1000-meters by Tucker and colleagues as “negative pacing” is correct. According to the classification given by Abiss & Laursen (2008), the most appropriated adjective for that pacing strategy is “parabolic” or, alternatively, “mixed” (according to the classification given by the Authors of the present study).
Paragraph 2, line 52: In line with the previous comment, the presence of a speed reserve detected by Kerhervé et al. (2015) suggests a parabolic/mixed rather than a positive pacing strategy.
Paragraph 2, line 55: Note that in the studies of Davies & Thompson (1979) and Millet et al. (2011) pacing has not been measured. Moreover, despite gradual increases in heart rate and VO2 have been observed in the study of Davies & Thompson (1986), I would use caution in interpreting those cardiovascular changes as a pacing strategy, since the treadmill speed was maintained constant for the whole duration of the test.
Paragraph 2, line 59: Angus & Waterhouse (2011) did not report differences in pacing variation between the three cohorts of running speeds assessed on ultramarathon distance.
Paragraph 9, lines 126-131: This part is copied-and-paste from the Methods section of the article of Kerhervé et al. (2015). This might be considered self-plagiarism and therefore be in contrast with the of the Journal policy on Publication Ethics. Please rewrite this part.
Raw data: Although the raw data have been made available as a part of Supplemental files and on the Figshare data repository, their reading results unclear and confusing since they have not been stacked and labelled in a clear manner.
The aims of the study are clear and the information about how GPS method has been used to assess pacing characteristics are described sufficiently in details. However:
Paragraph 7, line 107: A brief explanation on how to use the Vincenty formulae to measure point-to-point distances might facilitate the reading comprehension.
Paragraph 10, lines 150-153: What is the rationale and/or the scientific evidence used to determine these gradient ranges (LEV: -2.5 to 2.5%; UH: 2.5 to 100% and DH: -100 to -2.5%)?
Paragraph 14, lines 176-179/ Paragraph 22, lines 247-250: Please specify (if the Authors have access to those information) the reason of the pre-termination of the race (exhaustion, adverse weather conditions, etc) as it may have important implications on the interpretation of the results.
Paragraph 16, lines 205-213: There are no information about the characteristics of participants (training level/background) and about environmental conditions (temperature, wind, rainfall, etc.) occurred during the competition. Since these variables have been demonstrated to affect pacing strategies (Abiss & Laursen 2008; Hoffman 2014), their omission may have affected the power of data interpretation. Moreover, by uniting all the data according with their gradient range (LEV: -2.5 to 2.5%; UH: 2.5 to 100% and DH: -100 to -2.5%), how the Authors can exclude that the observed pacing variation at a given gradient has not been affected by the other interludes occurred within a given gradient range?
Paragraph 16, lines 213-216: Although using the average running speed as index of running performance permits to compare the two different groups (173-km and 103-km participants), have the Authors also tried to correlate the pacing characteristics with the finishing time in the two groups separately before concluding that there is no correlation between these characteristics and endurance performance?
Abstract, lines 25-28: This sentence is unclear: how positive pacing could have characterized all gradients if a speed reserve has been observed in LEV and UH?
Abstract, lines 28-30: This conclusion does not seem valid since significant correlations between overall performance and pacing characteristics have been previously observed in 161-km mountain ultramarathons (Hoffman 2014).
As I previously stated, this article is certainly interesting and it provides insight into some not well studied aspects of exercise science through a systematic analysis of pacing in a trail ultramarathon running event. However, there are some parts that definitely need to be corrected and/or rewritten in order to make it publishable.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.