I didn't read the log wrong, the expected result was -35.25767679620016 so the root cause of the problem must have been something else than I thought.

Is it possible for latency errors to mess up something as an expected result of -35.25767679620016 shouldn't be possible as the string only has addition, subtraction and multiplication with integers and the fix i implemented to let me pass the kata shouldn't have given me the result of -35.25767679620016 either?

To me it seems that a bug didn't let me pass the kata and another bug made me pass it. I believe that latency could have something to do with the wrong answers and errors (it did take over 500ms to pass the kata due to the inefficiency of my code)

Thank you for responding :)

for example one string that was to be evaluated was -(89) - (-17 * -71 * -(31)) + (92 - -((((24 - 77)))) + -49) which the eval function, the google search engine, wolfram alpha and my previous program evaluates to 37318 but was by the kata evaluated to be -35.25767679620016. From the 317 tests I got an error rate by about 2%.

Maybe i missed something, as u said it is highly unlikely that I'm the only one with this problem (if it is a problem). Anyway I hope this reply helps explain what i found to be wrong about the kata and I'd be happy to find out were I went wrong or how I missunderstood the kata

the test cases in python sometimes provide wrong answer due to the algorithm used to calculate the correct answers for the kata in python was wrong, if u put the string given in set test cases in a search engine or an eval function they would not have the same answers. I figured out what the kata did wrong and could then submit my kata which i altered to intentionally give the wrong answer so i could pass the kata which costed me a bit of time. Next time make sure your test cases are correct before publishing a kata as i bet multiple people atempting this one must be confused