Average Sudoku

Average Sudoku did not have a unique solution:

324 157 698
615 298 734
987 364 125
241 689 573
536 742 981
879 513 246
193 425 867
762 831 459
458 976 312

4 & 8 in Boxes 8 & 9 can be interchanged among themselves.

Please correct me if I am wrong.

Yes, it was on the XV grid... I had a 3 in r1c1 or r2c1 then 2 could not be in r1c1or r2c1, so I put the 2 r1c3 or r3c3, then I can not finish the grid....

And on average sudoku I thought there were 2 solutions, too...

Fred

it's my fault lost a "v" in vx sudoku, i'm very sorry to all.
the answer of average is
324157698
615298734
987364125
142689573
536742981
879513246
763825419
291436857
458971362
in the rule"The cell which value is the average of two cells aside marked with a line.", so if there's no line, it can't be average of two cells.

On the 7th line ... 2-5-8 would mean, that there is horizontal line over 5. So they can't be this way, only the other position is correct.

R7C7 can't be 8 because R7C6 = 5, and R7C5 = 2

oups you're right :) , I had chance to choose the right solution on this grid :bleh:

Fred

That was one tough set of sudokus !!
I liked the fact that the puzzles were all logically solvable (at least the ones that I was able to solve in the given duration)

There could have been more time allotted for both rounds.
I managed to solve only 3 in the variations round, solving a fourth consecutive just after the time was over.
It was a little sad that I wasn't able to try some of the cool sudoku like the Average Sudoku which has a neat set of logic involved.

It is very easy to miss an X or a V while making a puzzle like Sudoku XV, however it does distract the solvers and we need to formulate a process to avoid such errors in the mock tests in the future.
I will start another thread to discuss some of the possible ideas.

maybe because of the missing in vx sudoku, all the reply above are discussing variants, could you give me some reply for classic sudoku. like we metion in IB, all the puzzles are low level technology, 1-6 solvable by singles, 7 solvable by pointing, others need pair triplet and xwing, introduced in many japanese sudoku books. and we alse do some work on guessing part for 7-10.
let me show you the analysis by sudoku explainer(a sudoku analysis software)
Analyzing Sudoku #1
59 Hidden Single
Hardest technique: Hidden Single
Difficulty: 1.2

Analyzing Sudoku #2
57 Hidden Single
Hardest technique: Hidden Single
Difficulty: 1.5

Analyzing Sudoku #3
57 Hidden Single
Hardest technique: Hidden Single
Difficulty: 1.5

Analyzing Sudoku #4
51 Hidden Single
5 Direct Hidden Pair
1 Naked Single
Hardest technique: Naked Single
Difficulty: 2.3

Analyzing Sudoku #5
50 Hidden Single
3 Direct Hidden Pair
4 Naked Single
Hardest technique: Naked Single
Difficulty: 2.3

Analyzing Sudoku #6
50 Hidden Single
1 Direct Pointing
3 Direct Hidden Pair
3 Naked Single
Hardest technique: Naked Single
Difficulty: 2.3

Analyzing Sudoku #7
48 Hidden Single
1 Direct Pointing
6 Direct Hidden Pair
2 Direct Hidden Triplet
4 Pointing
Hardest technique: Pointing
Difficulty: 2.6

Analyzing Sudoku #8
51 Hidden Single
4 Direct Hidden Pair
2 Naked Single
4 Pointing
1 Claiming
2 Naked Pair
1 X-Wing
Hardest technique: X-Wing
Difficulty: 3.2

Analyzing Sudoku #9
53 Hidden Single
2 Direct Hidden Pair
1 Naked Single
1 Direct Hidden Triplet
6 Pointing
1 Claiming
1 Hidden Pair
1 Naked Triplet
Hardest technique: Naked Triplet
Difficulty: 3.6

Analyzing Sudoku #10
57 Hidden Single
1 Direct Hidden Pair
1 Direct Hidden Triplet
12 Pointing
2 Claiming
9 Naked Pair
Hardest technique: Naked Pair
Difficulty: 3.0

we know players like to get more score puzzles first, so we put the puzzle seems easier at last with high score. compare with your step and analysis, maybe you will have a new feeling with the ten puzzles, and i wonder someone can tell us your feeling, or you solve time for every puzzle, if all puzzles according with the theme metioned in IB, or something you want to tell us. thank you! :)

Thanks for correcting me. I thought that the converse might not be true, so I didn't check for that.

Thanks for the awesome analysis. Nice way to rate the classics.

I think all the techniques that you have used are absolutely acceptable in competitive sudokus. These techniques can be identified by participants without breaking their head too much or being tempted to resort to trial and error. :)

score / Results has been updated to show the split between Round1 and Round2.

Couple of observations
1. The average score in Round1 is much higher than average score in Round2. So, probably Round2 should have been longer.
2. Many players didn't attempt Round2. (so, either many players are just interested in Classics OR some players gave up after finding Round1 little tough)

Only Zafer completed Round1 with 4 minutes of bonus time. So really a good show from considering many other good players solved only 5-7 in Round1

As already noted, scores in this mock are very low, implying that the test became tougher than it was perhaps planned by the organisers. We can perhaps infer that two short 50 minute formats can work well in an offline competition, but it is not optimal for online contests.

Round 1: Two contestants completed all 10. So timing wise, it was planned ok. If we use the grades from scanraid, we arrive at the following:

1 119 moderate
2 126 moderate
3 101 moderate
4 82 moderate
5 106 moderate
6 84 moderate
7 154 tough
8 306 tough
9 166 tough
10 194 tough

I am not sure about the accuracy of this system but this does imply that the 20 pointer was tougher than the 40 and 50 pointers. And the 90 pointer was not tougher than the 70 pointer. In fact, classic 8 was attempted by the least number of contestants (another reason could be that it is difficult to reach to number 8, whether you start from 1 or 10). But overall, the points distibution was good as all "moderate"s had less points than "tough"s.

Round II: Consecutive sudoku was nice with logical steps needed to proceed. Thats the only one I can comment on since I got stuck in most others I tried, and solved diagonal the hard way. Here, we probably tried to fit in more sudokus than the alloted time. In 50 minutes, 5 or 6 sudokus (of the level in the mock) would have been perfect. The main regret is that most participants were unable to even try out a majority of sudokus.

Another reason for people skipping round 2 could be the rigid "20 minute gap" constraint between the rounds. If the two rounds were independent, we would have seen more scores in round 2.

I had some informatic problems during round 1 so I could try only the easier puzzles, which were very pleasant to solve. I think 50' was a good timing for this round.

Then I tried round 2 but did mistake on mistake and so I decided to stop playing and relax. :bleh:

I came back on the puzzles later, and my impression is that puzzles were globally hard, and some really tough. It would have been great to have the round last about one hour and a half.

Finally, the 20' break between the two rounds didn't bother me.

I would be a bit careful about relying on scanraid ratings for human solving. It doesn't seem to correspond too well with actual testing. For example, see Nick Baxter's remarks about the USSC puzzles here (and from direct experience, the 3rd puzzle of the third round was pretty much the toughest one of the bunch).

I agree with this, Scanraid's ratings are sometimes far away from what we could expect. Of course every software will sometimes deliver "strange" scores according to us, but Sudoku Explainer is generally considered more reliable. Anyway I didn't try the tough puzzles yet, so I can't tell about my own experience...