MAPE and sMAPE

In the theory section, we mention how sMAPE is known to treat overestimation and underestimation unequally.

Let's consider the following scenario. We have true targets, $Y$, and for simplicity, $Y = \{100\}$.

We have two models, $M_1$ and $M_2$, which produced the following predictions: $\hat{Y_1} = \{85\}$ and $\hat{Y_2} = \{ 115 \}$.

Calculate their sMAPE and answer the following question: which model showed greater performance based on the sMAPE? Your answer function looks like this:

$$\text{Answer}= \begin{cases} \text{sMAPE}_{1}, & \text{if }\ \text{M}_1\, \text{showed better performance} \\ \text{sMAPE}_2, & \text{otherwise} \end{cases}$$

Round the answer up to the third decimal. Use the modified sMAPE formula (the one that outputs in the $0-100\%$ range).