Is it a myth that having bottles with ice in the freezer improves performance? My dad told me to put bottles filled with water in the freezer and keep them there, with the purpose of giving it more "thermo-inertia" (the temperature changes slower to external influences) and this could make the freezer somehow (I don't understand how, if it does) save energy.
Now, is this reasonable? How much energy could I be saving?
Obviously, the drawback is a big space of the freezer is user for the bottles. Really, if it did save energy, maybe the manufacturer would have put a device himself.
IMPORTANT: This question is identical. But it's on engineering SE.
 A: I would like to qualify the answers offered by @Bort and @sammy gerbil based on crude estimates given below.
Consider a fridge with perfect thermal insulation. Initially it is filled with air at room temperature, $T_{room}$. Let $V_{fridge}$ be interior volume of fridge. If all the air inside the fridge is to be cooled down to some average temperature $T_{cold}<T_{room}$, then you must remove heat
\begin{align}
 Q=\rho_{air}V_{fridge}C_{p,air}(T_{room}-T_{cold})
\end{align}
Since fridge has perfect thermal insulation there is no loss, and $T_{cold}$ can be sustained indefinitely if you don't open the fridge. If you open the fridge $n$ times, then assuming the worst case scenario in that all the cold air inside the fridge is replaced by room air, you need to extract $(n+1)Q$ amount of heat.
Now consider another identical fridge, in which fraction of fridge volume, call it $\phi$, is replaced by water, initially everything at room temperature. Cooling down to $T_{cold}$ requires heat removal, 
\begin{align}
Q'=[\rho_{water}~\phi V_{fridge}C_{p,water}+\rho_{air}(1-\phi)V_{fridge}C_{p,air}](T_{room}-T_{cold})
\end{align}
If you open fridge $n$ times as before, assuming again that all air inside is replaced by room air while also assuming that temperature of water remains unchanged and equal to $T_{cold}$ always, we need to remove heat equal to $Q'+n(1-\phi)Q$.
So the ratio of the amount of heat to be removed in the two cases is
\begin{align}
\alpha\equiv\frac{Q'+n(1-\phi)Q}{(n+1)Q}
\end{align}
Let $n$ be the average number of times you open the fridge between replacement of water inside the fridge. 
For any given $\phi$, savings will result if
\begin{align}
\alpha & < 1 \\
\frac{Q'+n(1-\phi)Q}{(n+1)Q} & <1 \\
\frac{Q'}{Q}+n(1-\phi) & <(n+1) \\
\frac{Q'}{Q} & < 1+n\phi \\
n & >\frac{1}{\phi}(\frac{Q'}{Q}-1)
\end{align}
Since $\rho_{water}C_{p,water}\gg\rho_{air}C_{p,air}$, then if $1\geq \phi \gg 0$ we must have $\frac{Q'}{Q}\approx\frac{\rho_{water}C_{p,water}}{\rho_{air}C_{p,air}}\gg1$. Thus we need 
\begin{align}
n>\frac{1}{\phi}\frac{\rho_{water}C_{p,water}}{\rho_{air}C_{p,air}}=\frac{1}{\phi}\frac{10^3\times 4.2\times 10^3}{1.2\times 1.0\times 10^3}\sim \frac{1}{\phi}10^3
\end{align}
Therefore even in the best case scenario where $\phi\approx 1$, you need to keep the water inside the fridge long enough that you open the fridge thousands of times between replacements of water, for any power savings to occur. This happens because the initial cost of cooling down water is much greater than that for air.
A: Update
In view of Zero's excellent Answer, I should point out that my Answer ignores the question of energy efficiency and addresses only the issue of temperature stability. I agree with Zero that this practice is not likely to save energy. 
One way in which a full freezer might improve efficiency is by reducing the frequency with which the compressor switches on, which requires a large current. If the temperature is more stable, it will take longer to warm up to the the 'on' temperature, so there will be less cycling of the compressor. But there are other factors which might outweigh the significance of this one.
Original Answer
Much of the food you place in the freezer will have a high water content, so if you keep the freezer full of food there is no need for bottles of water. The food  provides the "thermo-inertia" which your Dad recommends, and minimises space for cold air which falls out and is replaced by warm moist air at room temperature when you open the freezer door. (Having a chest freezer avoids cold air falling out. So does keeping food in drawers.)  
[Note : Fridge manufacturers recommend not over-filling. Leaving space for cold air to circulate helps keep the temperature uniform.] 
But if you are not able to keep the freezer full, it is an excellent idea, for the reason your Dad gives. You should fill as much space as possible in this way, while trying to avoid having to remove frozen bottles of water, which will thaw and have to be re-frozen later.
Although filling the empty space with blocks of polystyrene (polystyrol) will save energy, avoid the warm air intake problem, and reduce conduction of heat into frozen food, it will also reduce conduction of heat out of food which is not yet frozen. This increases the time it takes for food to freeze, which increases deterioration because the food is at a higher temperature for longer. The polystyrene will provide almost no "thermo-inertia".
Other useful discussions :
Seasoned Advice SE : Does keeping a freezer full significantly help energy efficiency
Skeptics SE : Does keeping a fridge or freezer full improve its energy usage?
Physics SE : Does an empty refrigerator require more power to stay cold than a full one?
The Straight Dope : Does a refrigerator cool more efficiently when full?
