Thread: How to write a Mastermind game without using array and pointer?

Match (and remove) all the exact matches first, then look for near matches.

So after the exact scan, you're left with 1 2 3 0 and 4 5 6 0 to compare with one another for near matches (of which there is zero).

I still don't know what's wrong with my program, someone pls tell me

Your test for "near" will still have to run through 4 digits, since you are putting the digits back in the place where they were.

I'm only guessing what's wrong from looking at the code - and that looks wrong to me. It is very helpful in posts like this if you (the original poster) posts what the "failure" is [e.g. "It doesn't calculate 'exact' the right way, it comes up with X when Y is the right result for guesses 1 2 3 4 where the correct answer is 4 7 3 5"].

--
Mats

Yes, you could use a mask. A mask would be a bit per value in the array. Start out with 0xF (1111 in binary), and clear the bits (and with ~(1 << n) ) when you "use the digit".

--
Mats

Originally Posted by wtxwt

But I know nothing about mask, I even don't know how to start, is the "mask" method better than the original one?

What is the meaning of "0xF" and "clear the bits when I use that digit"?

I don't think you need a mask - using a "marker" digit like the zero value should work just fine.

You probably need to check the "newguess" extracted number for zero before you check if it's a match or not - otherwise you'll count the 0's as matches, which they are not.

Code:

		for (i = 1; i <= 4-exact; i++)
		{
			g = newguess % 10;
			newguess /= 10;
			for (j = 1; j <= 4-exact; j++)

You should remove the red code.

You will also need to adjust the inner loop to start over with the original numer, as you will certainly have it zero when it's run through one loop. When you find a near match, however, you need to zero that digit out, so that it doesn't count again for another number of same colour.

--
Mats

The highlighted part has already kicked all the 0s out, but I still can't be able to correct the repeated numbers problem. You just compile it and play it for a while and you know what the problem is.

Try to enter some combinations with repeated numbers or let the program chooses some repeated numbers and the problem appears. I don't know how to solve it.

It takes me enough time to write answers to questions here, without actually trying to compile someone elses code, and "playing with it" until I figure out what the problem is.

You still need to remove the "color" when it's been "used" for a near match.

--
Mats

Originally Posted by matsp

It takes me enough time to write answers to questions here, without actually trying to compile someone elses code, and "playing with it" until I figure out what the problem is.

You still need to remove the "color" when it's been "used" for a near match.

--
Mats

I don't know how, as the checking run for 16 turns, if I don't reset the number, the near match will just become "1" from the 5th check to 16th check, but if I reset the number, some combinations will be counted repeatedly, sometimes the near match can be calculated as more than 4

e.g.
Answer: 3 6 3 1
Guess: 6 6 1 3
Number of near match: 3
The "3" is counted for two times, therefore 3 near match is displayed

Yes, so you set colorreset for the first three to zero. You also need to break the inner loop so you don't check more digits with the digit you've already used.

By the way, to make it easier to test a specific case, why not (for now) ask for the answer as well as the guess - so you can _SET_ your answer to any combination you like. If you still want the "random" version, if you enter zero at the first position [and any 3 other numbers], it generates the random variety as the existing code does. That way, you don't have to "play until the particular combination comes up" - it could take quite some time to get three of one or four of one, which is probably some of the more difficult cases.

--
Mats

Does that work for you when there are two same digits in the "color"?

I would expect it not to work correctly for the example you gave above.

Why not just do the "colorreset" modification inside the if-statement. Something like this:

Code:

                                if (g==c && g!=0)
				{
					near++;
                                        colorreset = colorreset - c*pow(10,j-1);
                                        break;
				}

--
Mats