The application of complex training for the development of explosive power


The ever-increasing emphasis that is placed on athleticism and sporting success has led scientists to investigate numerous training methods that can have a positive effect on performance. One such method that has received significant attention is complex training (CT).

This method of alternating heavy and light resistances has the end goal of improving power output. In their recent review of complex training Docherty et al. (6) credit Verhoshansky with early work in this field as far back as 1973. Although this was one of the first publications on the topic, one suspects that the Soviets (and possibly other eastern bloc countries) may have been using complex loading as a training tool for some time.

Researchers that have found CT to be beneficial credit a post-activation potentiation (PAP) as the major physiological factor. Docherty et al. (6) explain that the explosive capability of a muscle is enhanced after it has undergone maximal (or near maximal) contractions.

Although some studies have not found any benefit from performing this type of training, the majority of research has supported it’s application as a tool to enhance expression of muscular power and explosiveness. However as with most relatively new training techniques, there is a need for more long-term studies. More work also needs to be done to determine the optimal combination of training variables for different sports and those with varying training ages.

Definitions of CT

There is still some debate on exactly what constitutes CT. Fleck and Konter (11) quoted Verhoshansky’s simplistic definition as a series of exercises formed in succession with a goal of improving one physical characteristic. He went on to say that these exercises where designed to improve “explosiveness.” Examples of Verhoshansky’s complexes include:

  • Back squats with depth jumps
  • Kettle bell jumps with standing long jumps
  • Push off and knee lift with 30m sprints.

Ebben and Watts (10) completed a review of literature on CT in 1998 and expanded on Verhoshansky’s explanation. They defined CT as alternating “biomechanically comparable high-load weight training and plyometric exercises in the same workout” (p 18). The example of CT cited in this study was bench press and medicine ball power drop. Interestingly, the same authors state that at the time only one study had examined CT although it was widely practiced.

In 2002 Duthie et al. (7) described CT as “various sets of groups/complexes of exercises performed in a manner in which several sets of a heavy resistance exercise are followed by sets of a lighter resistance exercise” (p 530). These authors also mention the term “contrast loading” and define this as “the use of exercises of contrasting loads, that is, alternating heavy and light exercises set for set” (p 530). Of all the articles referenced in the current review, this was the only paper to mention contrast loading suggesting that more research needs to be done comparing these two training methods. It is possible that some authors may have indeed been investigating contrast training while referring to it as CT.

Docherty et al. (6) define CT similarly to Ebben and Watts (10) as “the execution of a resistance-training exercise using a heavy load (1-5RM) followed relatively quickly by the execution of a biomechanically similar plyometric exercise” (p 52). According to Duthie et al. (7) this definition is referring to contrast training!

In recent times it appears as though the invent of contrast training is the source of some confusion amoung scientists investigating the CT phenomenon. As further research is conducted this issue will no doubt be clarified.


Verhoshansky is acknowledged throughout the literature as being one of the first authors to publish on CT. In 1986 a paper written by Fleck and Konter (11) outline Verhoshansky’s explanation of the CT phenomenon:

“Professor Verhoshansky used the example of the perception of lifting a half-full can of water when you think it’s full. The excitability of the central nervous system responds in such a way that the water literally flies in the air because of the force applied. It is because if the body thinks it has to do more heavy work, so it remembers what is necessary to lift the full can and reacts accordingly” (p 66).

This “fooling” of the nervous system has been expanded upon in recent times to offer 2 proposed mechanisms for post activation potentiation (PAP). Docherty et al. (6) explain the first of these theories as enhanced “motor-neuron pool excitability” (p 53). Specifically, a greater neural effect is the result of any/all of the following:

  • Better motor-unit recruitment
  • Enhanced motor-unit synchronisation
  • Greater central input to the motor neuron
  • Decrease in presynaptic inhibition.

The same authors also propose that PAP may be produced by local muscle changes such as phosphorylation of the myosin light chain. This process is explained by heavy exercise increasing the amount of Ca2+ in the sarcoplasmic reticulum, and the sensitivity of the myofilaments to Ca2+. Essentially more Ca2+ at a cellular level enables more ATP to be produced, which in turn enhances power production. The authors of this study do not comment on whether they believe these local mechanisms are trainable or simply transient alterations.

Ebben and Watts (10) completed a review of complex training in 1998 and offered several explanations for PAP, listing the following possible factors

  • Neuromuscular
  • Hormonal
  • Metabolic
  • Myogenic

However only neuromuscular mechanisms are discussed in any detail, which may mean that the other factors were either not well understood at the time and/or were not supported in the literature. These authors also mention that the fatigue associated with heavy weight training may increase motor-unit recruitment during subsequent plyometric exercise. At this point it is important to note that more recent literature (6) tends to emphasise neural stimulation rather than fatigue when referring to PAP.

Ebben and Watts (10) also cite a study by Fees in 1997 that attributes PAP to the reciprocal inhibition around a joint caused by agonist stimulation. That is, a heavy bench press minimises any restrictive contraction of the rhomboids (antagonist) for the subsequent plyometric exercise. However it is known that this inhibition occurs within the first month of training for inexperienced lifters anyway. Whether this effect can be maximised in experienced trainers through CT has not yet been determined.

McBride et al. (18) and Gourgoulis et al. (12) also support the neural activation theories proposed by Docherty et al. (6). The former expand by proposing that CT training may increase neurotransmitter release in afferent nerves. They also reviewed several articles indicating that fast-twitch (FT) dominant muscles produce greater potentiation.

Dan Baker has published prolifically on the topic of CT and he mentions some alternate mechanisms to explain PAP. His 2003 paper (1) mentioned the series elastic component of the musculo-tendinous unit as a possible contributor. He commented that a resistance of 65% of 1repetition maximum (1RM) would favourably increase stiffness but heavier resistances (85-90% 1RM) would not be optimal. However one would wonder if 65% of 1RM would be sufficient to achieve enough neural stimulation that is the main producer of PAP as outlined in articles reviewed thus far (6,10,18,12). Training age and experience with CT would no doubt be factors that would influence the optimal intensity to achieve optimal PAP.

Baker (1) also suggests that the golgi-tendon organ (GTO) and Renshaw cell activation may be reduced as a result of a heavy stimulus applied to a muscle. Inhibition of these structures would prove advantageous considering their role in monitoring and limiting maximal motor-unit activation as a protective mechanism. However as has been mentioned, the reduction of feedback from these inhibitory structures can be seen in the first month of training in novices. Whether this inhibition is enhanced to a greater degree through CT has yet to be confirmed.

To summarise, the primary mechanisms for PAP appear to occur at a neural level and relate to a reduction of inhibition and enhanced motor unit excitability. Most authors acknowledge that local muscle factors may contribute to PAP but the literature is not extensive in comparison to neural factors.

Upper body and lower body studies

Quite a number of studies have been done investigating the effect of the squat exercise on subsequent vertical jump performance (7,11,12,14,15,17-19,21,22). Obviously the investigation of this “pair” of exercises is to do with their biomechanical similarity, a prerequisite that was outlined by Verhoshansky in 1973 (11).

Several studies reviewed concluded that CT enhanced subsequent expression of lower-body power (7,12,17,21,22). Duthie at al. (7) used 11 women to compare complex and contrast training methods through squats and jump squats. They found that contrast training did increase power output (although this was non-significant) but only for those with high strength levels. Gourgoulis et al. (12) published similar findings through their investigation of vertical jump ability. Subjects with greater maximal strength improved vertical jump by 4.01% after performing 5 x 2 half squats from 20-80% of 1RM. Those with lower strength levels only improved by 0.42%. However the authors in this study didn’t specify the training age of participants, merely stating that they were “physically active.”

McBride et al. (17) conducted an 8-week training program on 26 subjects who performed jump squat training at either 30% or 80% of 1RM. They found that the 30% group improved peak power and peak velocity at all intensities (30, 55 and 80% of 1RM). This group also improved 1RM and 20m-sprint time indicating a transfer to a sport-specific activity. Interestingly the group that trained at 80% of 1RM improved peak power and peak velocity at higher intensities (55 and 80% of 1RM) and increased 1RM, however 20m sprint times were significantly slower. This then suggests that to transfer gym-based power to a sport specific activity (such as a 20m sprint) lighter resistances should be used. However it should be pointed out that this study did not employ CT or contrast training, making comparison with other studies cited in this review difficult.

Similarly Tricoli et al. (21) investigated the effect of olympic style weightlifting exercises compared to plyometric training. They found that the olympic lifts were more effective at improving squat jump, 10m sprint and vertical jump. The subjects in this study were all physical education students who underwent 3 months of lower body specific training prior to the study and all were experienced weight trainers.

Although neither of these studies (17 and 21) used CT, their results indicate that a stimulus greater than plyometrics but less than 80% of 1RM is required to achieve a positive transfer to field-based activities such as sprinting.

Smilios et al. (22) found a short-term increase in countermovement jump occurred when intensities of 30% and 60% of 1RM was applied in the form of either a heavy squat or jump squat. Their subjects all had a training age of 2-3 years and participated in sports requiring explosiveness such as basketball, volleyball and soccer. The fact that the improvements were quite similar for vastly differing stimuli (heavy squat v’s jump squat and 30% v’s 60% of 1RM) suggests that for experienced trainers, a wide range of variables will provide a subsequent enhancement of power. Other investigators (7,12) also discovered that the most significant power gains occurred in those with higher strength levels.

Upper-body CT has not received as much attention as lower-body but some interesting findings exist nonetheless. In 2003 Baker (1) found that a set of bench press performed at 65% of 1RM improved bench press throw by 4.5%. In 2005 Baker and Newton (4) investigated the effect of an agonist-antagonist complex series. That is, subjects were tested on bench press throws but the experimental group performed bench pulls (antagonist) between tests. The experimental group improved by 4.7% in post-testing. In both of Baker’s studies (1,4) subjects were professional rugby league players who were experienced in strength and power training methods.

However not all studies have found a positive effect with upper-body CT. Hrysomallis et al. (13) found that 5 reps of 5RM bench press did not improve power produced from an explosive push up on a force platform. Subjects had an average training age of 3.1 years. Ebben et al. (9) used EMG and kinetic analysis to conclude that a set of 3-5RM bench press did not increase ground reaction force or EMG when performing subsequent medicine ball throws. This study used basketball players experienced in weights and plyometric training.

Brandenburg (5) found that a set of bench press at varying intensities (100%, 75% or 50% of 5RM) did not have a positive effect on bench press throws. This study only used 8 subjects, some of which had limited exposure to power training (no subjects had performed bench press throws previously). The author also only used average power as a measure over 3 throws (not peak power), and indicates that this may have influenced the results.

It should be noted that all of these studies used high intensity interventions (5RM) with subjects not experienced with strength/power training (with the exception of Ebben et al.). Therefore one must consider that fatigue may have still been evident when the post-testing was conducted. The variable of recovery will be discussed in greater detail in the next section.

Analysis of both upper and lower body CT studies reveal that lower-body CT appears to be more effective at improving power expression, although more research has been done in this area. It also appears that subjects need to posses a sound strength base and training age to demonstrate an increase in power production.

Suggested training variables

Throughout the work done on CT, the manipulation of training variables appears to have a significant impact on the magnitude of PAP and resultant increase in power production. Of particular importance appear to be the variables of training intensity and recovery.

Exercise mode is not a very controversial discussion point, as most authors have advocated the use of biomechanically similar activities for CT. The most common examples being bench press-bench throw and squat-squat jump. The theory being that PAP is more pronounced if the heavy exercise stimulates the same (or similar) neural pathways to the power exercise. However Baker (4) experimented with the use of a heavy antagonistic exercise (bench pulls) and found a significant increase in bench throw power (4.7%).

This then suggests that the contrasting activities do not need to be as biomechanically similar as has been previously suggested. Of course more work of this nature needs to be done even though the same author found that a “typical” contrast using bench press – bench throws yielded a similar power increase of 4.5%. Perhaps with trained individuals, the variables of intensity and recovery are more important than the mode of exercise?

Intensity of the “heavy” exercise is more varied in the literature ranging from 30% of 1RM (22) to 90% of 1RM (18). The use of 5RM is common (5,14,13,15,19), however a landmark study by Gullich and Schmidtbleicher [cited in Docherty et al. (6)] found that 3-5 maximum voluntary isometric contractions (MVIC) was “sufficient to increase explosive force in the upper and lower extremities and could be used to enhance performance and training” (p 53). However there has not been much subsequent support in the literature for the use of MVIC’s.

As was mentioned previously, those athletes with higher initial strength levels appear to gain more benefit from CT. Perhaps initial strength levels relate to the ideal intensity that can be applied to gain maximal PAP? Baker has conducted his work with elite level Rugby League players and has advocated the use of intensities around 50-65% of 1RM. Smilios (22) found that for lower-level (regional) athletes, at least 60% of 1RM should be applied for PAP.

Those studies involving athletes that did not find a positive response from CT used higher intensities of 3-5RM (7,9,14). This then suggests that lighter intensities may produce a more positive response in athletic populations.

Recovery is another variable that has a significant impact on the success of CT. PAP does operate in a certain “window” so if a further stimulus is applied too soon, fatigue will be evident. However too much recovery will result in the individual missing the window of opportunity, as the potentiation effect will have subsided. Certainly more long-term studies are necessary to determine if trained individuals can sustain this PAP effect for longer.

When comparing the literature, a wide range of rest intervals have been trialled from 10 seconds up to 5 minutes. With his elite power-trained athletes, Baker (1,4) found that 3 minutes rest between strength and power exercises revealed that potentiation was still present. Smilios (22) also found a positive effect with 3 minutes recovery between strength sets (only 1 minute rest between contrast sets). Several other authors applied 3 minute rest intervals (13, 15) but found no positive effect using a CT protocol. However both of these studies were conducted on non-elite subjects with small sample sizes (12 and 8 respectively). In fact Jones and Lees (15) admit that greater power output was demonstrated by their experimental group, even though the difference was not significant.

A number of studies have used 5-minute rest intervals (7,9,12,19) with varying results. Duthie et al. (7) and Gourgoulis et al. (12) found that those with higher strength levels showed greater PAP. However Scott and Docherty (19) and Ebben at al. (9) did not show any evidence of PAP with 5 minutes recovery. It should be noted that the former study (19) did not use athletes as subjects. Brandenburg (5) used 4-minute recovery intervals and found no evidence of PAP, however his subjects were only recreationally trained and did not allow use of the stretch shortening cycle during testing.

Jensen and Ebben (14) investigated a number of recovery intervals from 10 seconds to 4 minutes. They found no positive response to CT but do speculate as to whether results would have been different with a recovery period greater than 4 minutes. It was noted that a 10 second recovery period had a negative effect on subsequent power output.

There is no doubt that the degree of potentiation is related to individual factors such as strength levels and training history with CT protocols. Therefore a trial and error approach may be required to determine the optimal rest interval for different athletes across different sports. Looking at the literature, 3-5 minutes of recovery does appear reasonable for most athletic populations.

Volume of the intervention is another important variable that may influence the degree of PAP. Those studies that examined higher-volume (more than 1 set) strength exercises all demonstrated improved power output (7,12,22). However 2 of these studies (7,12) commented that a more pronounced effect occurred with stronger athletes. Fleck and Konter (11) also commented that Verhoshansky utilised 2 sets of a strength exercise to get a positive response, and it may be assumed that he was working with elite athletes who possessed greater strength levels.

Many studies (1,4,5,9,13,14,15,18,19) have examined a lower-volume (1 set) of strength exercises, and reported mixed results. A number of studies did not find any improvement in subsequent power production after 1 set of a heavy strength exercise (5,9,13,14,15,19), which may suggest that athletes may need more than 1 set to achieve the neural stimulation associated with PAP. However 3 studies (1,4,18) found that 1 set was sufficient to achieve a positive response, and all of these examined gridiron or rugby league athletes.

Therefore the jury still appears to be out as to whether a single-set or multiple sets are necessary to achieve PAP. Due to the inverse relationship that exists between volume and intensity, one may assume that CT protocols using a single-set of strength exercises would do so at a greater percentage of 1RM.

Periodisation of CT

If the coach has decided to implement CT for the physical preparation of his/her athletes, the next decision is when to use this form of training. Most of the studies that did not find a positive result with CT (5,9,14,15), commented that this form of training had no adverse effect on subsequent power output and could be used as a means of maximising available training time. Therefore it appears that, at worst, CT could effectively be used during phases of the season where training efficiency is important, such as the competitive phase.

Ebben et al. (8) and Simenz et al. (20) conducted surveys of major league baseball and national basketball association strength and conditioning coaches. Understandably coaches were not willing to discuss the specifics of their training programs, but many did indicate that they used CT (7/21 coaches for baseball and 12/20 for basketball). Similarly, 8/21 baseball and 9/20 basketball coaches, reported that they used plyometric training year-round. From these statistics we cannot say that all of these coaches used CT throughout the year, but they do indicate that CT is implemented at different stages of the training year.

Traditional methods of periodisation suggest that power training should be emphasised when attaining a peak, which typically occur at the start of the competitive season and for finals. Therefore CT is most likely to be utilised during the specific-preparation phase and as part of the lead-up to end of season games in team sports. However individual sports that do not compete on such a consistent basis may incorporate CT more regularly as a part of their build-up for major events.


The evolution of training methods in recent times has resulted in increasing scientific investigation into methods such as CT. Proponents of this training method have stated that the main physiological response is an increase in PAP which enables increased power production for the plyometric activity. Although studies conducted thus far have not reached a consensus, it does appear that CT can play an important role in athletic training for increasing power production and/or maximising training efficiency.

More studies are required to determine the optimal variables for CT but in the meantime, coaches are encouraged to experiment with the key training variables of volume, intensity and recovery to develop a model that is suited to their sport and athletes. Strong consideration must also be given to the initial strength levels of the athlete/s and experience with power training methods. These two factors appear to be prerequisites for significant increases in power production.

Suprisingly there has been minimal investigation on the effect of CT using biomechanically unrelated exercises. It would be interesting to see if lower body strength exercises would result in PAP in the upper body and vice versa. Baker has already demonstrated that the use of an antagonist strength exercise can augment power production (4) for the bench press throw. Such work should provide a strong basis for scientists wishing to further investigate this phenomenon.


  1. Baker, D., (2003) Acute effect of alternating heavy and light resistances on power output during upper-body complex power training. Journal of Strength and Conditioning Research 17 (3) p493-497.
  2. Baker, D., (2003) Acute negative effect of a hypertrophy-oriented training bout on subsequent upper-body power output. Journal of Strength and Conditioning Research 17 (3) p527-530.
  3. Baker, D. and Newton, R.U., (2005) Methods to increase the effectiveness of maximal power training for the upper body. Strength and Conditioning Journal 27 (6) p24-32.
  4. Baker, D. and Newton, R.U. (2005) Acute effect on power output of alternating an agonist and antagonist muscle exercise during complex training. Journal of Strength and Conditioning Research 19 (1) p202-205.
  5. Brandenburg, J. (2005) The acute effects of prior dynamic resistance exercise using different loads on subsequent upper-body explosive performance in resistance-trained men. Journal of Strength and Conditioning Research 19 (2) p427-432.
  6. Docherty, D., Robbins, D. and Hodgson, M. (2004) Complex training revisited: a review of its current status as a viable training approach. Strength and Conditioning Journal 26 (6) p52-57.
  7. Duthie, G.M., Young, W.B. and Aitken, D.A. (2002) The acute effects of heavy loads on jump squat performance: an evaluation of the complex and contrast methods of power development. Journal of Strength and Conditioning Research 16 (4) p530-538.
  8. Ebben, W.B., Hintz, M.J. and Simenz, C.J. (2005) Strength and conditioning practices of major league baseball strength and conditioning coaches. Journal of Strength and Conditioning Research 19 (3) p538-546.
  9. Ebben, W.P., Jensen, R.L. and Blackard, D.O. (2000) Electromyographic and kinetic analysis of complex training variables. Journal of Strength and Conditioning Research 14 (4) p451-456.
  10. Ebben, W.P. and Watts, P.B. (1998) A review of combined weight training and plyometric training modes: complex training. Strength and Conditioning Journal 20 (5) p18-27.
  11. Fleck, S. and Konter, K. (1986) Complex training. Strength and Conditioning Journal 8 (5) p66-68.
  12. Gourgoulis, V., Aggeloussis, N., Kasimatis, P., Mavromatis, G. and Garas, A. (2003) Effect of a submaximal half squats warm-up program on vertical jumping ability. Journal of Strength and Conditioning Research 17 (2) p342-344.
  13. Hrysomallis, C. and Kidgell, D. (2001) Effect of heavy dynamic resistance exercise on acute upper-body power. Journal of Strength and Conditioning Research 15 (4) p426-430.
  14. Jensen, R.L. and Ebben, W.P. (2003) Kinetic analysis of complex training rest interval on vertical jump performance. Journal of Strength and Conditioning Research 17 (2) p345-349.
  15. Jones, P. and Lees, A. (2003) A biomechanical analysis of the acute effects of complex training using lower limb exercises. Journal of Strength and Conditioning Research 17 (4) p694-700.
  16. Kotzamanidis, C., Chatzopoulos, D., Michailidis, C., Papaiakovou, G. and Patikas, D. (2005) The effect of a combined high-intensity strength and speed training program on the running and jumping ability of soccer players. Journal of Strength and Conditioning Research 19 (2) p369-375.
  17. McBride, J.M., Triplett-McBride, T., Davie, A. and Newton, R.U. (2002) The effect of heavy v’s light-load jump squats on the development of strength, power and speed. Journal of Strength and Conditioning Research 16 (1) p75-82.
  18. McBride, J.M., Nimphius, S. and Erickson, T.M. (2005) The acute effects of heavy-load squats and loaded countermovement jumps on sprint performance. Journal of Strength and Conditioning Research 19 (4) p893-897.
  19. Scott, S.L. and Docherty, D. (2004) Acute effects of heavy preloading on vertical and horizontal jump performance. Journal of Strength and Conditioning Research 18 (2) p201-205.
  20. Simenz, C.J., Dugan, C.A. and Ebben, W.P. (2005) Strength and conditioning practices of national basketball association strength and conditioning coaches. Journal of Strength and Conditioning Research 19 (3) p495-504.
  21. Tricoli, V., Lamas, L., Carnevale, R. and Ugrinowitsch, C. (2005) Short-term effects on lower-body functional power development: weightlifting v’s vertical jump training programs. Journal of Strength and Conditioning Research 19 (2) p433-437.
  22. Smilios, I., Pilianidis, T., Sotiropoulos, K., Antonakis, M. and Tokmakidis, S.P. (2005) Short-term effects of selected exercise and load in contrast training on vertical jump performance. Journal of Strength and Conditioning Research 19 (1) p135-139.
CHRIS HOWELL PHOTOGRAPHY Wil Fleming working out at Force Fitness and Performance in Bloomington, Ind. Friday, Feb. 24, 2012.
Wil Fleming working out at Force Fitness and Performance in Bloomington, Ind. Friday, Feb. 24, 2012.