For most problems, the original problem statements found on the IPSC website and in the downloadable archives should contain all the information you need to solve them. But in some cases, we had to publish additional clarifications or corrections during the contest. This page collects these announcements.

Additionally, while the Training Area faithfully reproduces most of the unique rules and special problems used in the history of IPSC, some problems had to be modified to work in the new format. (If you find that some problem or special rule isn't working as it should, and it's not mentioned on this page, please tell us! It's probably a bug.)

No global announcements have been preserved. Perhaps there weren't any.

**Problem D (Delayed search):**
Some teams were unable to submit problem D due to a technical issue. The contest
was extended and the teams were given another chance to solve the problem. You
do not get this chance in the Training Area. The virtual contest takes 6 hours,
but you can only submit for the first 5 hours. For the last hour, you can merely
watch the other teams solve problem D and overtake you in the ranklist. Sorry!

**Problem G (Gets and Puts):**
The quine in G2 has to be non-empty.

**Problem H (Hard Question – P=NP?):**
The example input is incorrect. The first line of the input in the example
should be 4 2. (There are only two rectangles.)

**Problem Q (Quack Strikes Back):**
The einuq in Q2 has to be non-empty.

**Problem B (Bottom Coder):**
The example for problem B1 is incorrect. Use a *single letter Q* as the
separator.

**Problem C (Crane Operator):**
The second condition in the problem statement is slightly wrong. It should say:

2. NumbersAlso note that different boxes can have different values ofxandyare nonnegative integers chosen in such a way thatpand no previous box was assigned to the location_{i}≠M mod Np_{i}

**Problem J (Just For Fun):**
In the J1 puzzle "palindrome", *previous* means previous from the given
date, not previous from today.

**Problem C (Counting Subsequences):**
The input file c2.in is missing a blank line before the first test case.
(All the numbers in the file are correct.)

**Problem H (h4x0r t3h c0d3):**
There is no automatic judging script for subproblem H2. In 2006 we checked the
submissions by hand. (Or rather, we would have if there were any.) Therefore,
problem H2 cannot be sumitted in the Training Area.

**Problem G (Gotta Pick'em All):**
To clarify, you are not allowed to pick up three *distinct* cards
A, B, C such that A+B=C. For example, the correct number of cards
for N=4 is 3, and there are two ways to pick them: {1,2,4} and {2,3,4}.

**Problem H (Here-There):**
The input specification is wrong. It should say:

Each test case consists of five integersThis has been fixed in the online problem statement, but it still affects the downloadable archives and problem statements.D,H,_{c}T,_{c}H,_{r}T, specifying the degree of the board_{r}D, the coordinates of Here (H,_{c}H) and the coordinates of There (_{r}T,_{c}T)._{r}

**Problem I (Inventing Test Data):**
In the example input, the explanation should read "between nodes
**1** and 3", not "between nodes 2 and 3".
This has been fixed in the online problem statement, but it still affects the
downloadable archives and problem statements.

**Problem H (Hunt!):**
This problem is not yet available in the Training Area. The judging scripts are
a bit harder to port than for other problems. If you'd like to see it, tell us –
send us an e-mail to the address in the footer.

**Problem E (Eyjafjallajökull):**
A more detailed description of the coordinate system:

When you change the X coordinate, you move horizontally, when you change the Y coordinate, you move vertically.

On the wind matrix, (-2,-2) is the upper left corner.

X grows to the right, Y grows downwards, as is usual with discrete arrays.

**Problem F (Flawed Code):**
An early version of the problem statement incorrectly stated that Alla's
solution takes 10 cubes. The correct value is 15.

**Problem H (Hash):**
In IPSC 2010, every team was assigned a random 20-digit number called a
TIS (Team Identification String). In the Training Area version of this
problem, assume your TIS is "31415926535897932384". (A proper solution should
work for any 20-digit number.)

**Problem M (Mass psychoanalysis):**
While you can still gain bonus minutes in the Training Area, only submissions
made in the original contest count towards the average.

**Problem H (HQ0-9+-INCOMPUTABLE?!):**
In IPSC 2011, every team was assigned a random 20-digit number called a
TIS (Team Identification String). In the Training Area version of this
problem, assume your TIS is "31415926535897932384". (A proper solution should
work for any 20-digit number.)

**Problem L (Lame crypto) and M (My little puppy):**
These problems are not yet available in the Training Area. The judging scripts
for these problems are in a huge mess and would likely need a complete rewrite
to work. If you'd like to see them, tell us – send us an e-mail to the address
in the footer.

**Problem D (Data mania):**
Anything that is listed as a country (including "_no_specific_country_") is a
country. Anything that is listed as an institution is an institution.

**Problem H (Harvesting potatoes):**
The input specification is wrong. The constraints for the easy input are
actually "r,c ≤ 20", not "r,c ≤ 12".

No errata.

**Problem T (Two covers):**
The input specification is wrong. Both subproblems should contain
"1 ≤ k ≤ ℓ" instead of "1 ≤ k ≤ n".

**Problem C (Copier):**
Unfortunately, the downloadable archives do not contain the correct inputs for
C1. Please download `c1.in`

directly from the website.

**Problem F (Find a sudoku in π):**
The problem statement originally contained a small typo in the introduction.
The sequence that occurs at offset 5 is "9265", not "9256".

**Problem H (Hashsets):**
On our machine, `sizeof(size_t) == 8`

in C++.

**Problem J (Judging programs):**
In subproblem J2, only instances where start != end are valid. (The problem
statement should have stated "two distinct endpoints".) If a program only
fails when start == end, it is still considered good.

**Problem L (Let there be sound):**
Some teams are having problems with determining the exact value of a certain
large number in **L1**. As this was not intended to be a part of
the problem, we decided to help you out by letting you know that the number
ends in 99.

**General:**
You need to install Python 2 and the Pygame library for some IPSC 2015
problems. You can test your installation using
this package.

**Problem D (Dijkstra's Nightmare):**
Graphs that contain negative cycles will be rejected by the grader. This is consistent with the
Python version of the reference implementation.
(The C++ version of the reference implementation will actually terminate on such graphs because its
long longs will eventually overflow. Please pretend that this does not happen. ;) Thanks to Mykhailo
Granik for pointing this out.)

**Problem E (Enclosure):**
The Pygame client needs the file cat.png,
which is missing from the downloadable archives. Please download it separately.

**Problem H (Humble Captains):**
In the example annotation, we had a typo: "and their difference is 0" should obviously say "and
their difference is 1". This is now fixed (except in the downloadable archives).

**Problem U (Unusual Game Show):**
Clarification: The contestant *does* know the probability *p*.
On the other hand, when she plays the game she *doesn't* know whether Monty is
tired for that specific game.

**Problem C (Counting swaps):** The PDF version has a typo in the example - the
first line of the input should obviously be 3, not 2.

No errata.