As someone who understands statistics I do see the value of TA and understanding patterns but how accurate can it be when analyzing Bitcoin which is highly sensitive to overblown news that have nothing to do with price trends.

When you have external factors that have a large influence on the rise and drop of the price, I don't see how patterns that occurred in the past are good predictors for price movements in the present or future. Not saying that TA can't be accurate, I do think that properly used as one of many tools, it can help a trader but in this example that you are looking at Fibonacci resistance of weekly price I feel is the equivalent of saying that nobody went to the park last two Mondays, therefore people don't go to the park on Mondays while failing to account that it rained in those occasions.