Initially, Banister et al. (1975) proposed describing the relationship between training and performance by means of two antagonistic functions: Fitness as a positive and Fatigue as a negative response to training (FF-Model). In the recent past another antagonistic model, the PerformancePotential-Model (PerPot), has been developed by Perl (2001). In an endurance training study, model fit and prediction accuracy of the models were satisfactory (Pfeiffer, 2008). However, the PerPot showed better results. To compare the two models with regard to some critical considerations a study in the field of strength training was carried out. Methods: Two male and two female subjects volunteered for an 8-wk course of strength training (monolateral biceps-curls and leg extension). The training (input) was measured by force plate/sensor and quantified as force impulse (Ns). Before each training performance (output) was determined as peak power (N), generated in a Maximal Voluntary Contraction (MVC). Intraclass Correlation Coefficient (ICC) and mean relative deviation (RD, %) between modeled and real performances were calculated for each subject, exercise, and left/right extremity (16 data sets) to verify model adequacy. Based on the model parameters and the training load, the performances of the last two weeks were simulated individually. Prediction accuracy was estimated by mean relative deviation between predicted and real performances of the last (RDp1, %), and last two weeks (RDp2, %). Results: A significantly better average model adequacy was achieved by PerPot (ICC: M = .73, SD = .16; RD: M = 3.68, SD = 1.46) in comparison to the FF-Model (ICC: M = .67, SD = .28; RD: M = 4.29, SD = 1.96), ICC: t(15) = -2.2, p = .045; RD: t(15) = 3.3, p = .005. The prediction accuracy for the final one and two weeks was also higher for PerPot (RDp1: M = 3.93, SD = 2.86; RDp2: M = 4.67, SD = 2.93) compared with the FF-Model (RDp1: M = 5.97, SD = 3.53; RDp2: M = 5.96, SD = 3.98), RDp1: t(15) = 2.7, p = .015; RDp2: t(11) = 2.3, p = .040. Discussion: Our findings confirm better results for PerPot as reported by Pfeiffer (2008). They were not in line with Ganter et al. (2006), who in a field study (cycling) proposed a lower prediction quality for PerPot. Local delay optimizations suggest that in cases of lower model adequacy there are two different phases of time delay on training.
© Copyright 2009 14th annual Congress of the European College of Sport Science, Oslo/Norway, June 24-27, 2009, Book of Abstracts. Veröffentlicht von The Norwegian School of Sport Sciences. Alle Rechte vorbehalten.