Heart Rate TRIMP: what is the evidence?

03 FEBRUARY 2020

Measuring heart rate is one of the most popular ways of monitoring load in team sports. If the heart rate of players were high during a session, we assume that it was a physically challenging session and vice versa. Because of this very intuitive method of monitoring load, it is one of the most popular methods in team sports. However, for achieving positive training outcomes we are not only concerned with the intensity of a session. Ultimately, the interaction between the volume (i.e. duration) and intensity will determine our training stimulus and our training outcome. Several researchers have developed methods to determine this training stimulus based on heart rate data. In this blog, we will explain Heart Rate TRIMP methods and their differences.

The easiest way of combining both volume and intensity in one variable would be to multiply the average heart rate of a session by the duration. Eric Banister was the first one to come up with an adjusted version of this approach. He called this variable Training Impulse (TRIMP) and included the duration, average heart rate, and an exponentially weighted factor for the intensity of the session (see formula 1 in the footnote).

Even though Banister’s TRIMP can be used in endurance sports, this approach is hard to implement in intermittent sports like football and field hockey. In these sports, the average heart rate is not a good representation of the intensity of a session. The average heart rate will not reflect the times when the heart rate is near the maximum heart rate of a player (for example during repeated sprints). It also wouldn’t be a good indication of the time spent in the lower speed zones (in essence around 70% of the time in matches). For these reasons, the use of this approach will not give an accurate value of the load placed on the players in team sports.

Table 1: Edwards’ method

To better account for the intermittent nature of team sports, Edwards and Lucia both defined different heart rate zones. For Edwards’ method, the maximal heart rate of a player (i.e. 100%) is used to define five relative heart rate zones and their corresponding coefficient (see Table 1). Lucia based her heart rate zones on the heart rate of the player which corresponds to low, moderate and high intensity exercise (see Table 2), and allocated arbitrary coefficients to each of these zones.

For both methods, the time spent in each of these heart rate zones is multiplied by the corresponding arbitrary coefficient to determine the TRIMP score. The drawback of Lucia’s approach is that each player has to perform a maximal exercise test in a lab setting to determine the heart rate zones. To make matters worse, improvements (or decrements) in physical fitness will affect these zones. For the successful implementation of this approach, players need to perform these maximal exercise tests multiple times in a year. Hence, in most player monitoring systems Edward’s TRIMP is used to determine the Heart Rate TRIMP.

Table 2: Lucia’s method

Next to this practical issue of Lucia’s approach, both methods also have fundamental limitations. By assigning arbitrary coefficients to the different heart rate zones, the methods assume that each second spent in the highest heart rate zone is five times (Edwards) and three times (Lucia) as effective for training adaptations compared to the lowest heart rate zone. However, there is no evidence supporting these assumptions. Furthermore, by creating zones, the assumption is being made that the training adaptation is the same for the whole zone regardless of where in the zone the player is training.

For example, if a player’s heart rate is 71% of his maximal heart rate the time will be multiplied by 3. However, if the player’s heart rate is 79% of his maximum, the time will still be multiplied by 3. This is an absolute heart rate difference of >15 beats per minute, but we assume that the same training adaptations occur. This assumption is also questionable. Therefore, based on these two limitations we might question the validity of both methods.

So this leaves us with the questions which options are left to determine the training stimulus based on heart rate data. There is evidence to suggest that the time spend above 90% of the maximum heart rate is related to positive changes in physical fitness. Even though this method also assumes that every second >90% of the maximum heart rate is equally effective for training adaptations (independent of whether the heart rate is 91% or 97% of the maximum)it makes the least questionable assumptions. So, for now, if we want to determine the training stimulus of a session, we can conclude that choosing the variable which makes the least assumptions (in other words: using the least modified input variables) seem the most useful in team sports.

Heart rate is one of the most popular methods of monitoring players, due to its intuitive nature. However, in this blog we have also seen that the heart rate load variables data might not be as straight forward as we think they are. Multiplying the average heart rate of a session by the session duration, does not seem to be a valid method for determining the load on players in team sports. In the second place, by creating heart rate zones and assigning corresponding arbitrary coefficient to them, we are making assumptions that are highly questionable. And even if we only determine the time in each heart rate zone, we are already making an assumption.

So, in order to be able to work with these variables, we need to be aware of the assumptions that are being made and their limitations. If you know these aspects, you need to decide for yourself whether certain variables will help you get insights in the physical status of your players.

Overall, we can conclude that heart rate is able to give us a rough idea on the way players perceive the load. But each variable has its own limitations. To get a better picture of the load on our players, we need to monitor more than solely heart rate ( sprint distance, accelerations, total distance). This will give us more information about what caused a higher or lower TRIMP score than expected, and also, what should be adjusted to optimize the performance of our players.

1. Banister’s TRIMP= duration* HRaverage,exercise- HRrestHRmax-HRrest*0.64℮1.92*HRaverage,exercise