I can't understand what are they saying? What's accuracy in this case of analog computer? I mean 0.01% accuracy is 99.99% not accurate. This can't be real. Right?
Full text is taken from "inadequate accuracy" of Narsingh Deo "System simulation with digital computer". This is surprisingly a good book, I didn't expect this to be a source for learning as most books written by local authors used to be garbage.

In general, the result from a digital simulation is more accurate than that from an analog simulation. The accuracy of analog simulation depends on the accuracy of the components being used which can vary from 0.01% to 2%. When accuracy required of components is more than 0.1%, the cost of components increases rapidly. An accuracy of about 1% for a simulation of a system with modest complexity is considered good.

 

No, 0.01 percent accuracy means the result might differ from the true result by 1 part in 10000.

 

They should be saying “error” but since the measurement of accuracy is to indicated by the amount of error, they just used shorthand. It would be very confusing.

 

They should be saying “error” but since the measurement of accuracy is to indicated by the amount of error, they just used shorthand. It would be very confusing.

it looks like this is used a lot common in electronics and instrumentation engineering.

 

They should be saying “error” but since the measurement of accuracy is to indicated by the amount of error, they just used shorthand. It would be very confusing.

also, they're saying more accuracy is good(thus expensive). should not less accuracy be better?

 

No, 0.01 percent accuracy means the result might differ from the true result by 1 part in 10000.

so why're they saying mor accuracy is better(thus expensive components). should not less accuracy be better? because observed readings would not fluctuate a lot?

 

also, they're saying more accuracy is good(thus expensive). should not less accuracy be better?

No, accuracy is desirable. Accuracy is expressed by using the error figure by convention. It’s a little easier to read so maybe that’s why. Smaller error numbers mean better accuracy.

An accurately worded version (no pun intended (originally)) might read:

Accuracy: .02% Error

The error part is redundant to an experienced person.

 

so why're they saying mor accuracy is better(thus expensive components). should not less accuracy be better? because observed readings would not fluctuate a lot?

Think about what accuracy and precision mean.

 

Neither accuracy, nor precision, is necessarily the goal of simulation. Understanding may the more valuable goal.

 

Neither accuracy, nor precision, is necessarily the goal of simulation. Understanding may the more valuable goal.

With a digital computer you need to also think about resolution (with real-world I/O).

 

The first Analog Computer I built consisted of 3 Logarithmic Slide Potentiometers and a beat frequency oscillator. Essentially an electric slide rule which was only as accurate as the parts used and my ability to correctly read the logarithmic scale for the pot. I would say it was less accurate than my 10" K&E slide rule. Now if I'd had some high-quality precision accuracy logarithmic trimmer pots with a vernier adjustment knob that was properly calibrated...

 

should not less accuracy be better? because observed readings would not fluctuate a lot?

Accuracy has nothing to do with any fluctuations in the signal.
With less accuracy the fluctuations would be the same, just not as accurately shown.
You seem to be confusing accuracy with precision.