Liquid water (LW) existence in pure ice below 273 K has been a controversial aspect primarily because of the lack of experimental evidence. Recently, electron paramagnetic resonance (EPR) has been used to study deeply supercooled water in a rapidly frozen polycrystalline ice. The same technique can also be used to probe the presence of LW in polycrystalline ice that has formed through a more conventional, slow cooling one. In this context, the present study aims to emphasize that in case of an external probe involving techniques such as EPR, the results are influenced by the binary phase (BP) diagram of the probe-water system, which also predicts the existence of LW domains in ice, up to the eutectic point. Here we report the results of our such EPR spin-probe studies on water, which demonstrate that smaller the concentration of the probe stronger is the EPR evidence of liquid domains in polycrystalline ice. We used computer simulations based on stochastic Liouville theory to analyze the lineshapes of the EPR spectra. We show that the presence of the spin probe modifies the BP diagram of water, at very low concentrations of the spin probe. The spin probe thus acts, not like a passive reporter of the behavior of the solvent and its environment, but as an active impurity to influence the solvent. We show that there exists a lower critical concentration, below which BP diagram needs to be modified, by incorporating the effect of confinement of the spin probe. With this approach, we demonstrate that the observed EPR evidence of LW domains in ice can be accounted for by the modified BP diagram of the probe water system. The present work highlights the importance of taking cognizance of the possibility of spin probes affecting the host systems, when interpreting the EPR (or any other probe based spectroscopic) results of phase transitions of host, as its ignorance may lead to serious misinterpretations.