Monday, July 2, 2012

RSA is overrated? Part 4

RSA is overrated?
Part 4

I want to finish this article so I can go to vacation with ‘clear head’, and spend my time eating gyros, swimming and playing beach volleyball instead of thinking about RST/RSA/RSS.

I am afraid that this article got away from sole RST/RSA/RSS discussion. The things I warned against, besides trying to answer whether RST/RSA is overrated, are the following

  • Watch out for the mean/average values – pay more attention to variability of the data, distributions, zones and how the data evolves over time
  • Watch out for the velocity based time motion analysis (especially with absolute zones without relative ones) – pay more attention to acceleration and power expression, not only movement speed.
  • Watch out for the linear/mechanical logic, especially deducted from the cross-sectional studies (whose conclusions were based on mentioned flaws) – pay more attention to longitudinal studies and complex systems logic
Regarding the first bullet-point, David Tenney from Seattle Sounders recommended me a book Flaw of Averages by Sam Savage worth checking. When I checked the authors website, the following picture popped out: 

I just can’t agree more. 

Another feedback I received lately was by Martin Buchheit regarding intermittent critical power (iCP) concept. He actually did a study on it (Int J Sports Med. 2008 Apr;29(4):307-15), which I forgot to reference. You can check it HERE. The conclusion was that iCP is basically hopeless.

When it comes to velocity-based time motion analysis, I already pointed out to the two posts, but I will repeat this one more time and stop bitching on it for the rest of the article (please check THIS and THIS  posts). Instead I will focus on the last bullet-point: Watch out for the linear/mechanical logic.

The mentioned linear thinking can be depicted in the following way:

Let’s deal with the first link. Will RST improve RSA? I’ve covered this problem in my Troubles with RSA article, but I want to give you a heads up on the article/commentary by Martin Buchheit in  Sports Med. 2012 Feb 1;42(2):169-72 (check it HERE). Basically, Martin concluded the same think (but with more proof). Here are some of the sentences:

...In support of this, a speed/agility training programme targeting the development of COD speed more precisely, resulted in greater improvements in both linear sprinting speed and RS performance than RST.[5] Taken together, these data suggest that RST is likely effective at improving pre-planned COD speed, but not for developing linear speed qualities and RSA per se. Therefore, the suggestion of using ‘‘repeated-sprint training to improve sprint performance’’ (p. 750[1]) is misleading

...Again, this suggests that the greater improvement in RS performance following RST is more likely related to the training specificity of RST and testing, rather than to improvements in RSA per se. In support of this idea, the only HIT programme that was more effective than RST for improving RS performance included COD at high speeds

...Finally, if sessions targeting ‘isolated’ physiological capacities need to be implemented, it is important to consider the best predictors of RS performance. While it is clear that maximal sprinting speed is the strongest determinant,[1] the relative importance of other parameters such as maximal aerobic speed (MAS), VO2max, and on and off-O2 uptake (VO2) kinetics have been overlooked.[1]

...A faster MAS is associated with a lower relative exercise intensity during the between-sprint recovery periods and reduces the proportion of the anaerobic reserve[15] that is used during the RS exercise.[16] This likely results in a lower anaerobic system participation and diminished local peripheral physiological disturbance and, in turn, a better RSA. In a practical way, these data suggest that the development of these two key running speeds (maximal sprinting speed and MAS) should be targeted first when the goal is to develop RS performance in team sport players. All additional training strategies targeting specific physiological adaptations (e.g. buffer capacity or phosphocreatine resynthesis) are also obviously welcomed.[1]

...To conclude, a combination of all training methods is probably the best option in real practice.[1] I will always keep in my records the 20045 season of my formerHandball team (Se´lestat, First French League), when, as a strength and conditioning coach, I almost exclusively implemented RST during the preparatory phase. While there is obviously no direct link between the players’ physical capacities and the game outcomes, the team had the worst start to the season in years (possible interference phenomenon[17]). Interestingly, the team became much more successful during the second part of the season, when handball specific but ‘isolated’ speed and HIT training sessions were implemented.

Taken from Sports Med. 2012 Feb 1;42(2):169-72

When I was chatting with Håkan Andersson we touched upon this RST à RSA issue (click HERE to read the interview with Håkan). His opinion was to take context/athletes into account – soccer players (with couple of exceptions) are not even close to high level sprinters, and thus training aimed for high level sprinters might not be needed for soccer players. This included speed work with very long rest, etc. So, RST might kill two birds with the same rock. I am not sure if I completely agree with this – I am more in line with Martin Buchheit that possible the combination of the methods might be the best option. I am not sure if I touched upon this, but I am pretty sure I warned against blindly applying training methods and planning/programming from power/endurance sports for mixed sports in THIS article.  To keep the long story short, in my mind the story of improving RSA goes like this:

Of course, this logic is valid only if improvement in RSA is related to improvement in Game Performance (and if that actually matter for the final outcome). I will come to this ‘flawed’ logic of mine later. 

When it comes to the logic that improving RSA or any other physical preparedness quality will improve game performance, most of our proof is based on the cross-sectional studies and correlations. In the letter to the editors (Eur J Appl Physiol. 2011 Sep;111(9):2387-9), Martin Buchheit  (yes, the hero of this article) and Alberto Mendez-Villanueva were critiquing the series of studies on YoYo test by a group of authors (Bradley et al.) in which the concluded that YoYo was sensitive to predict match-related physical performance in soccer. I won’t go in much detail here (make sure to check the article HERE) besides quoting the title (Physical Capacity-Match Physical Performance relationship in soccer: simply, more complex) and the following sentence: 

...Thus, the fact that players occupying different positions have similar physical capacities (i.e., Yo-Yo IE2 test) despite the marked differences in match physical performances suggest that the game’s tactical requirements rather than players’ physical capacity might be more important in determining on-field players’ activity patterns

Speaking from my own experience, we had GK that covered more distance in YoYo Intermittent Recovery Test Level 1 than couple of fullback and middle fielders. 

Thus, it is hard (impossible?) to conclude that players with better physical qualities (MAS, YoYo, etc) will perform better physically (more HIA, RSS and even ball contacts, etc) in a game (without taking game tactical constraints of the positions played into account), and it is even harder (more impossible?) to conclude that improving physical qualities will improve physical game performance, especially out from cross-sectional data. 

The first longitudinal study to measure effects of improvement/decrements in physical capacities on game performance that is going to be published in Int J Sports Med is Repeated High-Speed Activities during Youth Soccer Games in Relation to Changes in Maximal Sprinting and Aerobic Speeds by Martin Buchheit, B. Simpson and A. Mendez-Villanueva.  You can read the slides from the presentation at 3rd World Conference on Science and Soccer, Ghent, Belgium HERE

The advantages of this study is that it compared improvements/decrement in MAS (Vam-Eval) and Vmax (MSS – Maximum Sprinting Speed) on the occurrence and nature of repeated-sprint sequences, but assessed in two ways: absolute (defined as at least 1sec long and  above 19 km/h with maximum of 60sec recovery) [RHSS] and relative (defined as at least 1sec long and  above 61% of individual MSS with maximum of 60sec recovery) [RSS].

The disadvantages were that it involved only MAS and MSS tests (no RSA), the long delay between sprints in RSS (longer than the other studies, which used below 20sec), 1-Hz GPS unit and it involved junior players (14.5 year old players). Anyway, even with that ‘flaws’ this study is unique in it’s longitudinal study.
Also, improvements/decrements in MAS and MSS were judged as greater or lower than ½ of coefficient of variation (CV) of each test (which they assessed previously). This is great statistical method, along with using Effect Sizes, which they did.  

I won’t go into deep analysis of this study, but I will copy-paste some of the interesting sentences:

...It was recently highlighted in highly-trained young soccer players that the relationship between on-fi eld running performance and physical capacities is essentially position-dependent [ 7 , 30 , 31 ] . In fact, for some positions, the magnitude of the correlations between match running performance and physical capacities is very large (e. g., strikers), while for others, very low and non-significant (e. g., centre-backs). This is probably related to the fact that tactical duties associated with playing as a defender limit running activities. In contrast, attacking players have more space and opportunity to express and use their full physical potential

... We examined, in highly-trained young soccer players, whether substantial changes in physical capacities (i. e., MSS and estimated MAS) can impact repeated high-speed efforts (i. e., the occurrence and nature of repeated high-speed and sprint sequences) during international club matches. With data limited to 1-Hz GPS analyses, the main results are as follows:

 1) a substantial increase in either MSS or V Vam-Eval was at least likely associated with a greater RHSS occurrence, while the magnitude of this increase was position-dependent,

 2) a substantial increase in either MSS or V Vam-Eval was likely associated with a lower RSS occurrence during games, but for some playing positions only,

3) despite the limited available data, substantial reductions in either MSS or V Vam-Eval were associated with maintained or even small-to-moderate increases in the occurrence of both RHSS and RSS, and finally

4) the effects of substantial changes in physical capacities on the number of high-speed efforts and sprints per RHSS and RSS were clearly position dependent.

...Not surprisingly, the players who became substantially faster (+4 to +7 % in MSS) and/or fitter (+7 to +13 % in VV am-Eval ) managed to perform a greater number of runs above 19 km.h – 1 (i. e., RHSS) during games (+5 to +44 %). Present
results are in accordance with previous longitudinal studies, where well-trained young soccer players [ 23 , 26] were shown to cover a greater distance at high intensities after a short training period leading to improved physical capacities. Taken together, these data show that, at least until a given fitness level is attained [ 5 ] , improving both MSS and MAS is likely beneficial for high intensity running performance. While there is obviously no direct link between match running performance and success in soccer [ 11 ] , a higher match running capacity might still enable more involvements with the ball and increase scoring opportunities[ 19 , 23] . Whether an increase in MSS can have a greater impact on the ability to repeat high-speed efforts than an increased V Vam-Eval is not easy to examine with present data and limited sample sizes. It is however worth noting that the associated changes in game activity were of similar magnitudes. In a recent study in 61 team sport players, both MSS and peak incremental test speed were shown to be the strongest determinants of repeated-sprint performance [ 6] ; MSS showed however larger association with mean repeated-sprint time than peak incremental test speed. In this regard, both MSS and V Vam-Eval can theoretically impact high-intensity, intermittent
running performance, but in different ways

...In accordance with our initial hypothesis, these results show that changes in physical fitness and match running performance are not necessarily matched, and that playing positions can affect these relationships.  In other words, this suggests that the observed improvements in running performance during games may be more related to tactical and strategic factors than physical fitness per se . This idea is further supported by the fact that the players displaying a substantial decrement in V Vam-Eval performed more RHSS per games (i. e., full-backs and centre-backs) Similarly, midfielders showing a decrement in MSS managed to maintain their number of RHSS per game. Since a majority of the other players became fitter and performed more high-intensity running during the second series of games, we can hypothesize that the players showing impairments in physical capacities had to increase their work rate during games to match those of their teammates. This suggests that fitness level was an unlikely limiting factor to their high-intensity match running activity.

The older players, while being fitter, performed less sprints and less RSS than their less fi t (and younger) counterparts. Taken together, these data confirm that in relative terms, the faster/fitter players perform less RSS than the slower/less fi t players[ 31 ] and that repeated-sprint activity is more likely affected by game technical/tactical demands rather than by physical fitness itself [ 31 ] . In support of this, we also observed in the present study an increased occurrence of RSS in players presenting impairments in physical capacities (e. g., +20 % of RSS despite a
substantial 2 % decrease in MSS for midfielders). Because of their reduced locomotor speeds, these players probably had to use a greater proportion of their maximal sprinting capacity to match the running demands of their teammates (which translated into a greater number of individual ‘sprints’ as defined in the present study)

... It is however worth noting that, although poor physical fitness might not directly limit match running performance (as discussed above), it may increase relative match running demands [ 31 ] . This can affect decision making, passing ability [ 36 ] or fatigue development during the end of a game.

... To summarize, the present data shows for the first time that the changes in RSS
occurrence are independent of those in physical capacities, and that playing positions can affect these relationships.

... To conclude, with data limited to 1-Hz GPS analyses, the present results show that substantial improvements in both maximal sprinting and aerobic speeds can be beneficial to (absolute) high-intensity match running performance. However, changes in match running performance did not necessarily match those in physical capacities, i. e., the magnitude of the changes in match running activity was likely lower than that of physical capacities. Additionally, players displaying impairments in their physical capacities managed to maintain or even increase their high-intensity match running activities, and the changes in match running activity in relation to those in physical capacities
were position-dependent. These results confirm that, at least for some positions, physical fitness might not directly limit high- intensity match running performance. Game tactical and strategic requirements can affect on-field players’ activity patterns independently (at least partially) of players’ physical capacities [ 7 , 28 , 31 , 33 ] . These results also emphasize our lack of understanding of how “maximal physical fitness performance” influences on-field running performance, and off er a cautionary tale of viewing the physical fitness – match running performance relationship in soccer in a simplistic manner [ 28 ] . Further study should now examine the optimal position-specific training strategies and determine at which extent improvements in physical capacities are still beneficial for match running performance [ 5 ] and, in turn, overall match performance and match outcomes [ 19 ] .

Things are “simply, more complex”. And I will try to explain one more factor why I think they are even MORE complex. 

This study reported that players who showed decreases in MSS/MAS (physical capacities) were able to maintain HRSS/RSS (even improve amount RSS, which is based on MSS) and the authors hypothesized that those players had to increase their work rate during games to match those of their teammates. I wonder were these players more ‘fatigue’ after the game, more sore, demanding more easy training in practices, missed more practices, had more minor injuries?

This brings me to one concept that is along the lines with Stephen Covey’s Production and Capacity to Produce:

To be effective, one must find the proper balance between actually producing and improving one's capability to produce. Covey illustrates this point with the fable of the goose and the golden egg.

In the fable, a poor farmer's goose began laying a solid gold egg every day, and the farmer soon became rich. He also became greedy and figured that the goose must have many golden eggs within her. In order to obtain all of the eggs immediately, he killed the goose. Upon cutting it open he discovered that it was not full of golden eggs. The lesson is that if one attempts to maximize immediate production with no regard to the production capability, the capability will be lost. Effectiveness is a function of both production and the capacity to produce.

Downloaded from HERE

Even if we forget about the flaws of most of the studies (velocity-based time motion, cross-study design, reporting mean values, etc) can we conclude that something is insignificant and not worth training for if there is no statistical significance to the game performance and the final outcome? 

In the mentioned study by DiSalvo et al. they showed that higher ranked teams perform less HIA than lower ranked teams. In the study by Carling et al. they showed that RSS does not occur so much in a game and they concluded “that the low frequency of repeated high-intensity bouts observed in the present team suggests that this specific fitness component (RSA) might not play as crucial a role in elite match performance as commonly believed “. In the study by Buchheit et al. authors showed that the players that showed significant decrease in MSS or MAS were able to maintain HRSS and RSS activities in the game.

All those data suggest that RSA is over-rated.

Let’s pretend for a minute that if the future studies show that most physical qualities have low influence on physical game performance and that physical game performance have low influence on final game/season outcome? Would we then conclude that physical preparation training is insignificant for the soccer? Be it RSA, MAS, MSS, COD and training aimed at improving them?

I think we are missing one crucial aspect here. We are focused on production, while really forgetting about capacity of production. We want to kill the goose and take the golden eggs.

One thing that studies report (or not even report) and we don’t actually pay attention to are subject drop-outs. Well, in sport studies there are a lot of drop-outs and I would really love to see meta-analysis study of drop-outs. Most of the drop-outs were because of illness or injury (or club transfer). I would be more than interested in characteristic (both physical and game performance ) of those drop-outs.

Another thing I would love to see in the (future) studies is how these drop-outs (and injuries or just unavailability) of the players influences the season outcomes (make sure to check John Orchard article in Br J Sports Med 2009;43:963-965; full text available HERE, along with this ARTICLE)

My point here i, that we are forgetting player availability/durability during the long season and grueling sessions.

Even if improving physical preparedness (which one?) doesn’t improve game performance and final outcome is not related to game performance, what it might improve is player availability/durability and that might affect the season outcome to the higher degree than being able to improve RSS in the game. Yet, I need a proof for this. Anyway, I warned about “productivity~capacity of production” complementary pair.

Preparing for a single game is pretty easy, but preparing for 10 months long season with 1-3 games played per week, with regular practices and travel is way different story. Having a high work capacity is of utmost importance. The work capacity is usually defined by VO2max level or any other aerobic ability. Although might be related, work capacity is way more  - it is the ability to perform and recover from frequent specific work/training/playing. Smart practices and hard recovery is very important. In my mind work capacity is capacity of production, and great players are those with stable high level of performance over a long period of time. Even if their VO2max is low. I think there are huge genetic influences on work capacity, besides training. Without expanding this discussion to talent~practice complementary pair (make sure to check What Makes Champions? in Br J Sports Med. 2012 Jun;46(8):555-61 HERE; along with this ARTICLE that concluded: “Psychology plays an important role in training, competition, tolerance of pain and motivation. However, the role of genetic variation in determining psychological state and responses remains poorly understood; only recently have specific genes been implicated in motivational behaviour and maintenance of exercise. Thyroid hormone receptors exist within the brain and influence both neurogenesis and behaviour. With the current state of knowledge, the field of genetic influences on sports performance remains in its infancy, despite over a decade of research.”)  I might say that you could have the greatest engine in the world, but if you have lousy breaks and chassis you cannot do much with such an engine. Some players are just prone to injuries and some are very robust. How can we influence this with training is beyond me. And, no I don’t think it has to do with the newest super-functional, vibrational, unstable corrective exercise. This rather has to do with individualizing training loads and implementing smart player rotation and recovery strategies.

My final point is that we are forgetting capacity of production (player availability/durability) and more studies need to take this into equation.

The thing I also noticed is that there is a fine line between de-training and under-recovering (over-training) during the in-season. The common pattern is that players start to feel tired after the games. The coach thinks they might be training too much, and that wise strategy would be to decrease training load. They feel even more tired. He decrease load even more in hope they will re-fresh. They feel even more tired. This is the thing that were confusing me over time, a lot. Still is.

This might be related to finding by Buchheit et al. that even the guys that showed decrement in MAS/MSS were able to maintain RHSS and RSS, but that might make them work harder (that’s why we need time-motion analysis based on relative speeds/abilities). You decrease load, they de-train even more and the game (even if they don’t show lowered HIA/RSS – they might even increase due the bad tactics and technique due decrease practice  time and getting out of the groove) makes them more tired. Then the coach decide that they are de-training and he increases the load. Guess what? They are even more tired and sore. Their work capacity was lowered and the new workloads take time to get used to. I’ve noticed this situation numerous times and solving it is not easy. It reminds me about the “sharpening the saw” principle by Stephen Covey. 

Stephen Covey tells the story of a man who was walking through a forest when he came across a frustrated lumberjack.

The lumberjack was trying to cut down a tree with and was swearing and cursing as he laboured in vain.

“What’s the problem?”  The man asked.

“My saw’s blunt and won’t cut the tree properly.”  The lumberjack responded.

“Why don’t you just sharpen it?”

“Because then I would have to stop sawing.”  Said the lumberjack.

“But if you sharpened your saw, you could cut more efficiently and effectively than before.”

“But I don’t have time to stop!” The lumberjack retorted, getting more frustrated.

The man shook his head and kept on walking, leaving the lumberjack to his pointless frustration.

Downloaded from HERE

The problem might have been arisen because the coach wanted to peak for the long season, or because he was unable to continue sharpening the saw and individualizing workloads. Or maybe all of this is psychological and related to the grinding the season and routine trainings and training sessions/locations. This is still beyond me. 

Anyway, I am getting outside of the scope of this article. I want to finish and pack for the vacation. I am going to leave you with more complex graph I am working on and that touches on most of the things we were discussing here. Please note that this is work in progress.

Finally, this series are over and I can go with clear and calm mind to vacation. I hope you have enjoyed and learned something along the way.  Watch the big picture and continue sharpening the saw! 

Heading to Zakynthos!!!


  1. great job Mladen....greetings from Thessaloniki...

  2. This comment has been removed by a blog administrator.

  3. Hi Shyam,

    Thanks for kind words. I think there is a way to subscribe using email, but I am not sure how :) Maybe following using the Join button?


  4. Hi, probably our entry may be off topic but anyways, I have been surfing around your blog and it looks very professional. It’s obvious you know your topic and you

    appear fervent about it. I’m developing a fresh blog plus I’m struggling to make it look good, as well as offer the best quality content. I have learned much at your

    web site and also I anticipate alot more articles and will be coming back soon. Thanks you.

    Function Point Estimation Training