How to accurately calculate the time of creating a software product?

it's about the same what to learn to accurately predict the weather months in advance.

There are a good mantra:

taski 4 hours we do 4 hour tasks in 8 hours we do for 12, taski 16 hours a we 24, tasks for 40 hours we do not never.

Just try to do the decomposition of the problem into smaller pieces. And domniate the risk ratios (in your case - 4). Over time, this ratio will decrease and you will more precisely predict the timing.

In theory, theory and practice are the same. But in practice...
For a very long time to try to build a time Manager of a large level to play with the formulas to take into account the contributions of force majeure and will never become Wang. But life itself tells you the solution. I like using this method and never failed ever. You should draw up a plan to increase the period of PI. And to admit that despite your genius and performance - so be it. This is the psychology that draws us. Even the most detailed plan with dozens of subtasks does not include such items as: lunch, sleep, I got sick, I got drunk, cat have kitten, mother-in-law arrived, etc.. Assessing the task you assume 100% employment and the highest focus. Multiply the time on PI it's a guaranteed forecast (financial expenses this applies similarly).

well then you're in your x's take from the ceiling , why are you surprised that it does not coincide with the actual?
systems of project evaluation a bunch read about at least one.

There is such a thing called focus factor. Focus factor shows, what share of the average time spent by the worker to do useful work. In the best cases, the value of this coefficient reaches 0.7 mm, but can be considered a good result if the focus factor level of 0.6. In average commands, he is close to 0.55 in (hence "the principle of multiplying by 2").

It is very important to take into account when planning work. So, if the execution of a task requires 4 "ideal" hours, when the contractor no distractions when the focus factor of 0.6 it will spend on this task not 4, and almost 7 hours. Feel the difference?

In this case, each team member focus factor my. And if the timing is critical, you need in planning to consider the average value in the hospital, but the value of a specific performing specific tasks. But no one does.

Only the experience.
If you want something fundamental, then read Frederick Brooks "the Mythical man-month"

If the customer has experienced, he will multiply by 3, but your task is to approach the factor of 1.5 in the evaluation.
It is also worth considering that require completely accurate, the customer is entitled, providing the only completely accurate description of the product, in practice, the customer changes features, design etc. in the development process and often counts on the same terms.

You might help the head 12 of the bestseller "The Deadline"in order to calculate the time of creation of the software product.

Here is a summary of the Chapter:

1. Size of each project.

2. Especially at first with the choice of units of measurement — if you later you have to work with real data, it's a start and abstract units.

3. Build complex metrics based on simple (those are easy to calculate in any software product).

4. Collect archival data to consider the labour productivity in the already completed projects.

5. Work on the calculation formulas of complex synthetic metrics as long as the results are not will most accurately reflect the attitude of the abstract units to the specified archive data in the scope of work.

6. Swipe through the entire archive database the trend line, which will show the expected amount of work in the form of ratio values of complex synthetic metrics.

7. Now for each new project will be sufficient to calculate the value of the synthetic metric and use it to determine the expected amount of work.

8. Don't forget the "noise" on the line performance and use it as an indicator in determining the allowable deviations from the overall trajectory.