There is slight possibility I'm interpreting like I'm 5, so don't take this as gospel and please anyone else issue corrections, also it's a gross oversimplifica

There is slight possibility I'm interpreting like I'm 5, so don't take this as gospel and please anyone else issue corrections, also it's a gross oversimplification but here goes: First a brief outline of the actual experiment. Rather than directly measure all of the processes that occur within a plant to effect photosynthesis (from photons hitting a leaf and water being drawn from the ground to the creation of some sort of molecule used for transporting energy in plant cells e.g. ATP) the researchers cited in the article created a mathematical model of part of the process that directly concerns chlorophyll: absorbing photons and transferring the energy to the next stage. They don't specify what the next stage would be, they simply modelled a network of nodes from input (representing the initial photon hitting), via nodes representing different stages of chlorophyll, to a generic output. The models assumed two things indicated by prior research: 1) there is an upper threshold of energy production beyond which deleterious back reactions occur i.e. chlorophyll make too much power do nasty thing to plant 2) there is a lower threshold of energy production beyond which eventual power output is inefficient (nonlinear decrease in output) because the rate of energy transfer out of the network [from the chlorophyll stages to the next part of photosynthesis] is fixed by electrochemical processes i.e. chlorophyll make too little power makes actual photosynthesis even worse Given those assumptions an ideal chlorophyll stage would operate between the thresholds. In a static environment, that is one in which the rate and state of photons entering remains constant, this could be easily achieved. In a noisy environment, when the light source fluctuates as a result of shadows etc., then a system that relies on one absorption rate will see a direct corresponding fluctuation in power output - fewer/more photons in means less/more energy out and importantly the rates of power in to power out would be directly linked. This noisy environment could cause a drop below / rise above the 2 thresholds above and therefore bad/inefficient operation. Given this noisy environment is inevitable an ideal chlorophyll stage would smooth out the fluctuations in input energy so the output doesn't spike in direct relation to input spikes. Think here of adding capacitors to smooth current fluctuations. This smoothing can be achieved by having two (or more - this is limited by biological/chemical realities here) different concurrent input sources that flow energy through the system at different rates given they both produce the same output. Krackers has more clearly explained this bit than I did initially - although I must add the differing energy production rates is my interpretation of what's implied by the authors because they don't explicitly explain. I have ignored here the internal noise of the system which the paper attributes to protein dynamics driving fluctuations of intermediate excitation energy transfer events for simplification and because it doesn't effect the main thrust.