Simulation of electronic circuit design

I have a question about how do electronical engineers usualy proceed when starting electronical circuit design.

Should we always simulate all our designs using softwares (Pspice, proteus, ltspice)?
If for instance, I have all my electronical circuit design on paper, should I simulate it before ordering parts/boards?

it up to you
there is a rule says One hour of studying reduces four hours of working

I learned the basics of electronic design then began to design circuits before computers were invented. I designed all my circuits properly and every one worked perfectly.
I design so that my circuits work well if the parts are low, typical or high spec but today students do not understand that a simulator only looks a "typical" specs then frequently a circuit that passes the simulation will not work because the parts are low or high. You usually cannot order parts with "typical" specs, you get whatever they have.

As a hobbyist/learner, I'd say that Ahmed is right, reading up first..., but also a hobby that involves reading a lot and the calculator before starting anything usually, otherwise it's ill-informed doing stupid things for nothing mostly (beginner's mistakes).

However you engineers prototype (breadboard, simulation, pen and paper and head) surely what matters is knowing a circuit will work correctly, or will work enough to fine tune the prototype before spending your or some-one else's money? I'm no good at maths, wish I was, but as all circuits can be reduced to numbers, and numbers don't lie, I would think reliable calculations over not always reliable simulations are more important...

I gave up on Simulations, years ago.
I have had many circuits that work in Real live, but don't simulate correctly or even at all.
And others the Simulate good, But Don't work the same when actually Built.

Maybe the Newer Simulation Programs are Better.

But I Love HANDS ON work.
First Design it on paper, Than Proto type it on a Breadboard, and when Happy with it, make a PCB and Build it.

I think a Simulation Program never reads the datasheet for a device and simply guesses at what should happen.

Simulating a circuit will resolve a lot of headaches when attempting to build it.

For example if building a regulator or a filter you can be sure that over any temp/tolerance variations your circuit will work as expected or within your own specs. Like Audioguru mentioned students, including I, weren't told to consider these process/temp/tol variations when doing labs and hws. Now I am working full-time and going back to school and learning about all that and end of life (EOL) degradation of devices. It can add up depending on the environments your design will be in.

Also, EMI is hard to simulate as a hobbyist... so that's a big read-up, especially in discrete designs when laying out on a PCB.

i would say simulate it before ordering parts. even when you start a design with calculation or theory, simulation is essential to predict EM performance because you are not designing your circuit using Maxwell equations! simulation technology have advanced in recent years. there are companies like modelithics which will provide simulation models for your lumped and active components for keysight ADS so your simulation will be very accurate and close to measurement. for EM simulation CST provide good tool. they have done alot of advancement for their tool in recent years to simulate EMI, EMC simulation apart from s-parameter simulation.

I very rarely use simulation.
The problem is either so simple simulation is pretty much a waste of time, or its so complex correct fully accurate modelling is not really possible.

Its good for some fairly straightforward tasks such as active and passive filter design. But you will find things such as chokes and transformers, real ones will behave very differently to pure perfect "mathematical" inductances.

Simulation often assumes many things, such as perfect noise free grounding, and no adverse coupling through a common power supply rail.
In some cases physical layout is absolutely critical to success, but a simulator will never tell you that.

Trying to fight ringing and oscillation in a fast high impedance voltage comparator can be quite daunting, but those types of frustrating problems never happen with simulation.

The usefulness of simulators has been proved. They offer more than they detract. Even though a simulator can make mistakes, we can make mistakes too.

My own electronics knowledge has grown as I developed my own homemade simulation program. It has limitations. I experiment with other simulators as well. They have their limitations. Some limitations are obvious. Some become apparent with use after a while.

Even though simulators can't do everything for us, they're a useful tool. The world of electronics has grown so broad that we benefit from using any tool that can further our understanding.

From the direction the thread friendly chat is taking, re to simulate or not to simulate, I wasn't saying simulators are no use, that would be silly. Every tool has its place when making something.
I imagine as the years go by it will become impossible to breadboard as the parts will be so small and BGA etc. that the only method will be simulation, or being Einstein with a calculator, which as one poster maybe sort of said - won't be the case for most of us.
Simulators are useful as they can show up a lot of things you may not see or think about, but they are also quite hard to use and get a meaningful result from if you don't set every single parameter individually or understand the virtues and limitations of a set of algorithms, which I certainly can't do, and as some-one else said: best laid plans of designers and then the reality shows up unexpected quirks.
Anyway, all tools have a place, and should make the job easier, as does the tool of real world analysis, but it's pointless saying one method is "better" than another as no toolkit is 1 screwdriver or 1 hammer.

A person showed a simulation of a portable preamp for an electret microphone. The preamp used one transistor.
The Simulation showed that everything was fine. The SIM software did not know that:
1) When the 9V battery was brand new the transistor was saturated and did not work.
2) When the battery was used a lot but still had some life then the transistor was cutoff and did not work.
3) When the battery voltage was near one of the above then the signal was rectified with bad distortion.
4) When the sound level was high then the transistor was very non-linear with terrible distortion but it was not clipping.