Towards the end of the book The Hungry Brain, by Guyenet, there's a reference to research where both the sleep deprived group and the control group are let loose in the city and allowed to eat whatever they wanted, conditioned on letting the researchers weight and classify it first.
On average, the sleep deprived group ate around 300 calories more.
(sorry I can't produce the original reference, I don't have the book with me)
The mechanism that drives this are the hormones ghrelin and leptin. Look up their action if you're interested.
Also, to your point, very few people spend their sleeping hours eating. You could eat 1500 calories for every meal and be very lithe - if you slept 23 hours a day. By contrast, it seems anecdotally that people who get less sleep tend to eat more food and more of it is unhealthy: late night Taco Bell, delivery when you're pulling an all-nighter, etc.
Best diet pill on the market? Unisom.
Not sure if this statement is sound, for the average person.
If you ate 4500 per day, spread across 3, 20 minute meals, and spent the remaining 23 hours sleeping... your energy expenditure would be minimal, whilst your energy intake large. The average person would gain weight rapidly. (300lb athletes perhaps not, but these are outliers).