Well, a quick update that doesn't mean too much. I checked the kill a watt this morning and over the past 35 hours the fridge has used exactly 1 kWh/day. This is a 6% reduction in energy consumption. However, with how much the power consumption seems to fluctuate, I am not comfortable saying that it will stay at 6% and am pretty sure it will change. I have still not swapped out the milk jugs yet, but I have added a juice container I had. Its more rectangular and is probably roughly 2.5 quarts.

As for Piwoslaw's ideas, I think they're good ones. I'd imagine if you have ice you'd want it on the top, if you are just using water the bottom would be better for the reasons you've mentioned. That does seem like an interesting test and perhaps I will do that next.

Well, I now have 95 hours of logging on the fridge and I'm looking at 1.04 kwh/day. Not real promising. I'm going to let it keep logging and will keep swapping the containers in and out. I am using 3 milk jugs and 1 juice (~2qt) container since I didn't have 4 milk jugs. The jugs have never fully thawed in the fridge before going back outside, however, they have turned to quite a bit of water.

I've been getting approximately 15% savings on my 1999 power-hungry fridge KWH (213 watts running, 3 watt standby per Killawatt) consumption by using a heavy duty digital timer to turn it off from 23:30 till 05:30 every night. I've logged the temp in the upper section of the fridge with a recording USB temperature stick. Normal temperature is 38+/-2F. With the fridge door closed and off for 6 hrs at night, the temp climbs to about 42. In the morning the fridge runs for a longer cycle than normal to get the temp back to normal setpoint. I've been doing this for 2 years now.

I would use Blue Ice blocks in place of water jug. They hold more cold, fit in smaller places, are made for long use and are common at thrift stores in my area. Plan to give it a try(though have no means to test it). Our frig never seems to have much empty space though.

Remembering what I can from a physical science class, Fridges are pretty much heatpumps. So wouldn't it make more sense to somehow connect the radiator to a thermal cool mass? The fact that the appliance tends to be butted up against a wall where heat can build up probably doesn't help too much. You want less of a temperature differential so it has to do less work.

But it also has e temp switch that turns it on above a selected temp. If it never reaches that temp it doesn't cycle. But yes a refer runs better in a cold space than a hot one.

Just a small update here. My wife turned on the ice maker a few days after I started the testing with the ice jugs which spiked energy usage . I have since restarted logging data with ice jugs.

I still have no solid info on this. The recent heat wave (it was 50F the other day) has left little ice to be had. The jugs in the fridge are now just water and the energy consumption is above 1.0 kwh/day.

If I had to guess, and its very unscientific, I'd say the jugs reduced my power usage by ~5%.

I've been thinking about adding a twist to this idea: instead of transporting heat from the house to the outdoors (through the refrigerator), reverse the process.
When the outdoor temperature (say, 12°C/54°F) is less than indoors (18°C/64°F), but more than in the fridge (3°C/37°F), place your jugs of water outside until they are warm, then put them in the fridge which will now pump the heat into the house. When the water has cooled, put them back outside and repeat.

This pretty much turns your fridge into a water/air heat pump, only the heat source (water) side is manually operated. Quite primitive, and probably not as efficient as a dedicated heat pump/minisplit, but since heat is being transported into the house, then the reduction of required heating (from other sources) will outweigh the increase of energy consumption by the refrigerator. Using the freezer (temperature around -20°C/-4°F) instead of the refrigerator would extract much more heat from the water.

That's an interesting idea, and you would be transporting heat into your house when you set 3°C jugs outside and brought 12°C jugs in. Is your fridge as an ASHP more efficient than your regular home heating equipment?

If you do the math, I think you'll find moving jugs of liquid water moves far too little energy to make much of a dent in your heating bills. The fact that the water changes phase when it's cold outside is what allows you to transport a meaningful amount of heat in the jugs.

I've thought about placing my fridge's condenser outside the thermal envelope of the house. However, that would deprive me of the heat generated by the fridge, and the heat that leaks from the kitchen into the fridge. My condenser would spend a fair amount of energy making icicles.