Many of the materials used in modern construction, like drywall, plaster, tile and concrete, don't burn. While they may break up as the building collapses around them, they will remain, and by studying them, the investigators can get a very good idea of where the fire started and how it spread. If a molotov cocktail were used for example, the splashes from the burning material would be evident on whatever concrete or drywall surfaces that remain.
Only in a lightly framed, all wood structure, with no plaster, concrete, or masonry to speak of, would all evidence be effaced, though of course the glass would remain. Such construction would be virtually alien in the United States.
Nonflammable material such as tile or concrete has to be scorched enough by the gas burning for the scorch pattern to remain despite massive amounts of water. Drywall sucks up water like crazy, making it hard to use. But if a structure fire is left to burn without human interference, it still won't destroy completely even things like wood. Especially because of collapse, areas will not burn completely. Sometimes a piece of paper will survive a structure fire, because it didn't reach sufficient temperature to ignite before being buried in an oxygen poor environment, such as underneath a wall. Complete structure fires are interesting because of how the fire travels. Sometimes entire walls are left untouched. Wild fires exhibit this issue even better.
The flash marks of a Molotov cocktail would be very evident if the fire is suppressed quickly. This is the case with any accelerent. But as the fire burns longer and longer, it does more destruction, and the suppression efforts are stronger, such as water and tearing down walls, increasing the destruction. After a significant length of time of burning, it becomes very difficult to determine the cause of a fire.
Given an hour to burn any house fire would totally destroy a normal house.
One method I heard of which leaves no forensic trace was to take a cigarette, start it going, fold the middle of the cigarette over into a pack of matches (so as it burns down it will touch off the matches. Then, leave the matchbook and cigarette down between the cushions on a sofa. The sofa goes up like a candle and it's enough to set pretty much anything ablaze and the small matchbook and cigarette are overlooked. Even if traces are found it's assumed that someone was smoking and started the fire accidentally.
There was a firebug in California who used this method for years. The only reason they caught him was that he was also a fire investigator and he just happened to be on hand at a coincidentally high number of fires that were out of his jurisdiction to lend a hand. They only busted him in the end by tailing him around and catching him in the act.
The cigarette in the matchbook is now old, but it never left no forensic evidence. It frequently leaves the bottom of the matchbook where it is all stapled together and the striker strip is located. Problem is when the technique was new, investigators didn't know what to look for. Further, although finding such a strip is indicative that the fire is arson, it doesn't help you one bit at finding out who started the fire.
Remember that evidence can indicate a crime happened, it can indicate a certain individual committed the crime, or it can indicate both. For example, video of van der Sloot with the victim indicates that he was with the victim, and nothing more. By itself, it is meaningless. The dead girl in a room indicates a crime happened by the manner of her death. But the dead body does not indicate who committed the crime. You have to add all the evidence together.
In the case of the cigarette/matchbook trick, it indicates a crime happened, but is highly unlikely to indicate who committed it.
If you live in the US or the UK and have a newish sofa, it won't burst easily into flames. All material such as mattresses, sofas, and chairs are required in those countries to be treated with flame retardant. There's at least one video showing the difference between being treated and not being treated.
The question isn't so much "Will there be any evidence?" as it is "Will the investigators find it?"
Levels of expertise and resources vary depending where you are. Not every police department is CSI savvy.
A Molotov cocktail would be pretty obvious arson. But if the firebug had instead tried to start an explosion near a gas line or a water heater, an overworked cop without a lot of arson/forensic training might miss the signs...
Most arson investigation is done initially by the fire department. Until you've been in a significant number of fires, you don't understand the nature of the beast. Fire is a living, breathing creature that operates in ways that are both predictable and unpredictable. Until you understand fire, you can't understand how to investigate it.
With arson investigation, one issue is when people develop new ways to start fires. It takes a while before the investigators catch up, although this is a problem in many types of crimes.
Arson investigation tends to be a volume business. Small police and fire departments don't have one. In New Hampshire, all arson investigation is done by two or three people. They do it for our larger cities as well as towns with populations in the three digits. It takes a very big city to warrant an arson investigator, even then, the total number is small.
Best of luck,
Jim Clark-Dawe