How accurate is the lichess engine/computer analysis?

I would like to know how accurate the computer analysis is. I have had some games a 60 move game for example. My opponent got 1,0,0 Im certain he didnt use an engine as it ended in a draw. I got 1,1,0 (inaccuracy, mistake,blunder) centipawn loss of 8. I dont think this is very accurate.

1. I don't see games like what you describe. There are no such games on this account. It is very hard to comment this anyway when you're using some abstract examples.
2. In long games both you and your opponent can easily get low centipawn loss if the game went mostly in a calm endgame with 0 or so computer assessment.
3. The power of analysis on lichess by one of the three best engines (Stockfish 8) is capable of beating Magnus Carlsen at standard OTB time controls with overwhelming probability. Any errors in analysis like too high rated position which is in fact an easy draw can happen because of some deep misunderstanding by SF 8 but in fact are not very common. They cannot affect centipawn loss very much.

the computer analysis is accurate.
your mistake is that you are making assumptions about "cheating" based on numbers that should be taken into context, like 8 centipawn loss can mean a lot of things (one side gets an overwhelming advantage very quickly, quiet game, luck, cheating..)

I'd like to mention that an engine evaluation is really just an opinion and is not either right or wrong, unless it gives a mate in n or 0.0 evaluation.

1. isn't 0.0 also an opinion?
2. it is an opinion based off rigorous coding and decades of development. Insofar as any opinion can be right, an engine's opinion is as right as you can get. Personally, I wouldn't even call it an opinion.

An opinion to me is referring the black side of a French defense because of x, y and z. A computer's 'opinion' is far more elaborate, exact and mathematically founded.

Of course, or else chess would be solved, but insofar as any opinion can be accurate, wouldn't that of an engine be 'the best' due to its evaluation process, as it were?

Back to the original question, while it is rare, it is completely possible to play a game where there are not many tactics to be found. Usually very defensive style play with little aggressive (forward moving pieces). More common is your opponent blunders badly, you find it and from there on out most of your moves are good moves. Same if you blunder for your opponent.

except the rule of fortress, any other positions are correct. Number at that position is useless

Insofar as any opinion can be right, an engine's opinion is as right as you can get
Nope, there are such things as "Freestyle chess" and "Advanced chess", where people analyze positions with the help of engines and get more accurate moves than engines would found themselves. Also in the opening theory it is often known much more about the relative quality of moves than an engine can tell you at low depth, and in addition, Lomonosov tablebases are more accurate in 7-men positions than modern chess engines even using 6-men Syzygy.