review notes

[ghamerly]

@jsannemo

Some things to fix before HiQ... ~I will add to these as I have time to do so and add a note when I'm done.~ I'm done.

• [ ] halloweenloot validation errors:

  ERROR in submissions: WA submission halloweenloot2_js.cpp (C++) got RTE [test case: test case secret/01-rnd-knapsack, CPU: 0.01s @ test case secret/01-rnd-knapsack]
  ERROR in submissions: WA submission halloweenloot3_js.cpp (C++) got RTE [test case: test case secret/03-rnd-knapsack, CPU: 0.00s @ test case secret/01-rnd-knapsack]
  WA submission halloweenloot4_js.cpp (C++) OK: WA [test case: test case sample/2, CPU: 0.00s @ test case sample/2]
  ERROR in submissions: WA submission halloweenloot5_js.cpp (C++) got RTE [test case: test case sample/1, CPU: 0.00s @ test case sample/1]

• please link to the source of the illustration images for:
  • [x] costumecontest
  • [x] halloweenloot
  • [x] isithalloween
  • [x] ministryofmagic
  • [x] witchdance (if used)
  • [x] wizarddance
• isithalloween
  • [x] s/halloween/Halloween/g
  • [x] the argument to \url should have a leading https://
• halloweenloot
  • [ ] s/lexicographically smallest/lexicographically earliest/
• ministryofmagic
  • [x] the sample data could use an explanation (I wouldn't be surprised to get a lot of clarification questions from weaker contestants)
  • [x] add a comma after \dots
• witchdance
  • [x] s/embarrasment/embarrassment/
  • [x] s/dancy/dance/
• wizarddance
  • [x] add a comma after \dots
  • [x] s/wizards should/wizard should/
  • [x] Typically a wizard is a male noun, so using she sounds off here.

[ghamerly]

I tried to spend more time thinking about the ministryofmagic sample, and I don't understand it. Probably that's because I'm a weaker contestant. But here's my thinking (and likely what someone else might do); what's wrong with it?

1. Initially there are 4 candidates: 1, 2, 3, 4
2. In the first round of voting, the 4 parties vote for candidate 1 (two votes) and candidate 4 (two votes). The unpopular candidates are 2 and 3, and 3 being the lowest-numbered is eliminated.
3. In the second round of voting, candidates 1, 2, and 4 remain. (Here is likely where I'm thinking incorrectly.) Since every party's preferences still have their highest-ranked candidate available, the voting goes the same way: candidates 1 and 4 each get two votes. Candidate 2 is eliminated.
4. In the third round of voting, the only remaining candidates are 1 and 4. The voting goes the same way (since each party's favorite is still available). Candidate 4 is eliminated (it is the lowest-number least-popular candidate). Thus, candidate 1 wins.

[jsannemo]

1 is lower than 4, so 1 is the lowest-numbered least-popular candidate

[jsannemo]

Updated some of the statement fixes. I don't agree re lexicographically smallest vs earliest: lexicographically smallest is by far more commonly used in statements.

[jsannemo]

Re my WA solutions: is this tested locally or on Kattis? WA vs RTE is notoriously difficult to determine between environments: on my computer they still get WA rather than RTEs. If you tell me what they get on the Kattis slave you install on I will move them to the correct places.

[jsannemo]

Added image sources too

[ghamerly]

1 is lower than 4, so 1 is the lowest-numbered least-popular candidate

Okay. Then my confusion was in being tired, but also because voting preferences indicate that the earliest (lowest) number is preferred (as in being the first in the permutation ranking 3, 1, 2, 4 indicates 1 is preferred), the rule for breaking ties is that being larger (higher numbered) is preferred. So that might be worth mentioning.

[ghamerly]

Updated some of the statement fixes. I don't agree re lexicographically smallest vs earliest: lexicographically smallest is by far more commonly used in statements.

Both are okay; I point it out because Google prefers "earliest" (on the web in general).

[ghamerly]

Re my WA solutions: is this tested locally or on Kattis? WA vs RTE is notoriously difficult to determine between environments: on my computer they still get WA rather than RTEs. If you tell me what they get on the Kattis slave you install on I will move them to the correct places.

The original was on my laptop; I just ran it on a kitten and got this:

Loading problem halloweenloot
Checking config
Checking statement
/home/hamerly/problem_installation/hiq-challenge-2018/problemset.cls exists, will not copy it -- in case of weirdness this is likely culprit
/home/hamerly/problem_installation/hiq-challenge-2018/problemset.cls exists, will not copy it -- in case of weirdness this is likely culprit
Checking validators
Checking graders
Checking data
Checking submissions
   AC submission halloweenloot_js.cpp (C++) OK: AC [CPU: 0.00s @ test case secret/10-rnd-knapsack-zeros]
   AC submission halloweenloot_mw.cpp (C++) OK: AC [CPU: 0.13s @ test case secret/11-allzeroa]
   AC submission sl.cpp (C++) OK: AC [CPU: 0.10s @ test case secret/07-rnd-knapsack-zeros]
   Slowest AC runtime: 0.126, setting timelim to 1 secs, safety margin to 2 secs
ERROR in submissions: WA submission halloweenloot2_js.cpp (C++) got RTE [test case: test case secret/01-rnd-knapsack, CPU: 0.00s @ test case secret/01-rnd-knapsack]
ERROR in submissions: WA submission halloweenloot3_js.cpp (C++) got RTE [test case: test case secret/03-rnd-knapsack, CPU: 0.00s @ test case secret/02-rnd-knapsack]
   WA submission halloweenloot4_js.cpp (C++) OK: WA [test case: test case sample/2, CPU: 0.00s @ test case sample/1]
ERROR in submissions: WA submission halloweenloot5_js.cpp (C++) got RTE [test case: test case sample/1, CPU: 0.00s @ test case sample/1]
   WA submission halloweenloot_js.cpp (C++) OK: WA [test case: test case secret/18-twostartingpoints, CPU: 0.00s @ test case secret/10-rnd-knapsack-zeros]
   RTE submission halloweenloot7_js.cpp (C++) OK: RTE [test case: test case secret/98-maxdiff2, CPU: 0.00s @ test case secret/10-rnd-knapsack-zeros]
   RTE submission halloweenloot8_js.cpp (C++) OK: RTE [test case: test case secret/98-maxdiff2, CPU: 0.00s @ test case secret/12-allzerob]
halloweenloot tested: 3 errors, 0 warnings

[jsannemo]

Thanks, will update!

Could you share TLs for the problems as run on the kitten so we can review them?

On Thu, Nov 1, 2018, 12:27 Greg Hamerly <notifications@github.com wrote:

Re my WA solutions: is this tested locally or on Kattis? WA vs RTE is notoriously difficult to determine between environments: on my computer they still get WA rather than RTEs. If you tell me what they get on the Kattis slave you install on I will move them to the correct places.

The original was on my laptop; I just ran it on a kitten and got this:

Loading problem halloweenloot Checking config Checking statement /home/hamerly/problem_installation/hiq-challenge-2018/problemset.cls exists, will not copy it -- in case of weirdness this is likely culprit /home/hamerly/problem_installation/hiq-challenge-2018/problemset.cls exists, will not copy it -- in case of weirdness this is likely culprit Checking validators Checking graders Checking data Checking submissions AC submission halloweenloot_js.cpp (C++) OK: AC [CPU: 0.00s @ test case secret/10-rnd-knapsack-zeros] AC submission halloweenloot_mw.cpp (C++) OK: AC [CPU: 0.13s @ test case secret/11-allzeroa] AC submission sl.cpp (C++) OK: AC [CPU: 0.10s @ test case secret/07-rnd-knapsack-zeros] Slowest AC runtime: 0.126, setting timelim to 1 secs, safety margin to 2 secs ERROR in submissions: WA submission halloweenloot2_js.cpp (C++) got RTE [test case: test case secret/01-rnd-knapsack, CPU: 0.00s @ test case secret/01-rnd-knapsack] ERROR in submissions: WA submission halloweenloot3_js.cpp (C++) got RTE [test case: test case secret/03-rnd-knapsack, CPU: 0.00s @ test case secret/02-rnd-knapsack] WA submission halloweenloot4_js.cpp (C++) OK: WA [test case: test case sample/2, CPU: 0.00s @ test case sample/1] ERROR in submissions: WA submission halloweenloot5_js.cpp (C++) got RTE [test case: test case sample/1, CPU: 0.00s @ test case sample/1] WA submission halloweenloot_js.cpp (C++) OK: WA [test case: test case secret/18-twostartingpoints, CPU: 0.00s @ test case secret/10-rnd-knapsack-zeros] RTE submission halloweenloot7_js.cpp (C++) OK: RTE [test case: test case secret/98-maxdiff2, CPU: 0.00s @ test case secret/10-rnd-knapsack-zeros] RTE submission halloweenloot8_js.cpp (C++) OK: RTE [test case: test case secret/98-maxdiff2, CPU: 0.00s @ test case secret/12-allzerob] halloweenloot tested: 3 errors, 0 warnings

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/jsannemo/hiq-challenge-2018/issues/1#issuecomment-435012200, or mute the thread https://github.com/notifications/unsubscribe-auth/AA6GXbzuFO97yXmFbwl8Ihnva4rnBywiks5uqtqMgaJpZM4YFYRS .

[ghamerly]

As run on a kitten:

Loading problem costumecontest
   Slowest AC runtime: 0.025, setting timelim to 1 secs, safety margin to 2 secs
Loading problem halloweenloot
   Slowest AC runtime: 0.127, setting timelim to 1 secs, safety margin to 2 secs
Loading problem isithalloween
   Slowest AC runtime: 0.024, setting timelim to 1 secs, safety margin to 2 secs
Loading problem ministryofmagic
   Slowest AC runtime: 0.436, setting timelim to 2 secs, safety margin to 4 secs
Loading problem witchdance
   Slowest AC runtime: -1.000, setting timelim to 1 secs, safety margin to 2 secs
Loading problem wizarddance
   Slowest AC runtime: 0.314, setting timelim to 2 secs, safety margin to 3 secs

[jsannemo]

LGTM. Uploaded moved solutions

[ghamerly]

@jsannemo are you ready for me to install the problems?

[jsannemo]

Yes! We skip the last one.

On Thu, Nov 1, 2018, 18:47 Greg Hamerly <notifications@github.com wrote:

@jsannemo https://github.com/jsannemo are you ready for me to install the problems?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/jsannemo/hiq-challenge-2018/issues/1#issuecomment-435121989, or mute the thread https://github.com/notifications/unsubscribe-auth/AA6GXbMbGLAzuf6Zc_oBdpvAtzs1yOmkks5uqzPLgaJpZM4YFYRS .

-- /Johan

[ghamerly]

Okay, the problems are installed. I installed the versions on the branch corresponding to this PR: #2 (because the first one I installed was sometimes giving 404s on the image due to its size).