rng_58's blog

World Finals 2017 statistics and my prediction
By rng_58, history, 8 years ago, In English

1. Tsinghua (Win: 40%, Gold: 90%, Medal: 100% after some rounding)

  • CF max: 3441, 3203, 2967 (agg = 3516)
  • Open Cup: 1, 1, 11, 9 (median = 5)
  • China Final: 1
  • MIPT: 1, 1, 1, 2, 4, 1, 4

2. Warsaw (Win: 15%, Gold: 75%, Medal: 95%)

  • CF max: 3122, 2962, 2826 (agg = 3237)
  • Open Cup: 2, 3 (median = 2.5)
  • Petrozavodsk: 4, 2, 4, 3, 2, 2, 1, 2

3. SPb ITMO (Win: 15%, Gold: 50%, Medal: 90%)

  • CF max: 2880, 2870, 2860 (agg = 3104)
  • Open Cup: 7, 3, 2, 1, 13, 1, 2, 1, 3, 7, 13, 8, 8, 12, 3, 2, 1, 12 (median = 3)
  • Petrozavodsk: 4, 1, 3, 1, 6, 5, 3, 7, 5
  • MIPT: 6, 2, 3, 1, 5, 2, 1

4. SPb SU (Win: 10%, Gold: 40%, Medal: 85%)

  • CF max: 3170, 2951, 2936 (agg = 3281)
  • Open Cup: 9, 21, 18, 5, 3, 4, 9, 9, 6, 1, 2, 2, 7, 1, 4, 23, 20, 4 (median = 5.5)
  • Petrozavodsk: 7, 3, 6, 5, 8, 1, 4, 5, 1
  • MIPT: 2, 4, 2, 3, 1, 10, 8

5. Seoul (Win: 5%, Gold: 40%, Medal: 85%)

  • CF max: 3174, 2919, 2760 (agg = 3250)
  • Open Cup: 4, failed, 4 (median = 4)

6. MIPT (Win: 5%, Gold: 30%, Medal: 75%)

  • CF max: 2928, 2757, 2756 (agg = 3063)
  • Open Cup: 16, 29, 27, 28, 7, 24, 17, 15, 9, 15, 17, 3, 9, 4, 10 (median = 15)
  • Petrozavodsk: 1, 5, 5, 5, 1, 6, 6, 1
  • MIPT: 3, 6, 5, 4, 6, 3

7. MIT (Win: 5%, Gold: 30%, Medal: 75%)

  • CF max: 2900, 2637, 2540 (agg = 2981)
  • Open Cup: Equivalent to 2.5 in NAIPC

8. KTH (Gold: 10%, Medal: 60%)

  • CF max: 2761, 2637, 2273 (agg = 2870)
  • Open Cup: 2, 17, failed, 7 (median = 12)
  • MIPT: 5, 5, 16, 7, 7, 2

9. Fudan (Gold: 5%, Medal: 55%)

  • CF max: 2543, 2367, 2327 (agg = 2667)
  • Open Cup: 6, 13, 9, 6, 13, 18 (median = 11)
  • China Final: 5
  • MIPT: 12, 3, 7, 15, 13, 4, 5

10. New South Wales (Gold: 5%, Medal: 50%)

  • CF max: 3073, 2429, 2006 (agg = 3082)
  • Open Cup: 6, 8, 16 (median = 8)
  • MIPT: 14, 9, 10, 16, 2, 5, 7

11. Tokyo (Medal: 50%)

  • CF max: 2762, 2694, 2424 (agg = 2903)
  • Open Cup: 29, 5, 10, 12, 14, 29 (median = 13)
  • MIPT: 11, 10, 6, 8, 15, 6, 6
  • Barcelona: 1, 1, 1, 2, 3, 1, 1

12. SJTU (Medal: 50%)

  • CF max: 2707, 2609, 2582 (agg = 2874)
  • Open Cup: 26, 3, 18, failed, 21, 12, 10, 20, 26, 14 (median = 19)
  • China Final: 2
  • MIPT: 8, 7, 4, 10, 3, 9, 10

I won't be surprised if the following teams get a medal:

13. NTU (Medal: 45%)

  • CF max: 2772, 2341, 2220 (agg = 2808)
  • Open Cup: 11, 15, 21, 10, 2 (median = 11)

14. Waterloo (Medal: 45%)

  • CF max: 2925, 2908, 2703 (agg = 3100)
  • Open Cup: Equivalent to 12.5 in NAIPC
  • Barcelona: 3, 2, 4, 1, 1, 3, 2

15. Perm SU (Medal: 40%)

  • CF max: 3037, 2850, 2578 (agg = 3127)
  • Open Cup: 6, 18, failed, 3, failed, 9, 18, 7, 24, 15, 14, 16, 16, 7, 21, 16, 21, 17 (median = 16)
  • Petrozavodsk: 8, 12, 10, 14, 10, 10, 2, 6, 6
  • MIPT: 4, 8, 11, 5, 20, 7, 9

16. Peking (Medal: 40%)

  • CF max: 2920, 2566, 2338 (agg = 2965)
  • Open Cup: 21, 3 (median = 12)
  • China Final: 6

17. Helsinki (Medal: 35%)

  • CF max: 2721, 2280, 1516 (agg = 2745)
  • Open Cup: 15, failed, 9, failed, 16, 4, 17, 11, 30, 18, 2, failed, 8 (median = 16)
  • Petrozavodsk: 3, 26, 25, 11, 11, 7, 10, 2, 9
  • MIPT: 10, 16, 14, 6, 23, 17, 3

18. BSUIR (Medal: 30%)

  • CF max: 2825, 2625, 2466 (agg = 2943)
  • MIPT: 7, 13, 8, 11, 9, 13, 12

===

  • Expected win of top 7 = 0.95
  • Expected gold of top 10 = 3.75
  • Expected medal of top 18 = 11.05

And I guess Tokyo Institute of Technology will be around 25th, Keio will be around 40th.

  • Vote: I like it
  • +365
  • Vote: I do not like it

»
8 years ago, # |
  Vote: I like it +22 Vote: I do not like it

Is there any mirror contest for ICPC World Finals?

→ Reply
»
8 years ago, # |
  Vote: I like it +21 Vote: I do not like it

IMHO, one of the best predictions, which I see for this year's final. Btw, high probability that Perm SU and SJTU reach top 12.

→ Reply
»
8 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Auto comment: topic has been updated by rng_58 (previous revision, new revision, compare).

→ Reply
»
8 years ago, # |
  Vote: I like it +39 Vote: I do not like it

How do you calculate the probability? :D

Btw, I'm very surprised when I see SpbSU ranked relatively low in this prediction and many other predictions:

  • 2 members of SpbSU won ACM ICPC WF last year.
  • This team also won NEERC this year.
  • Individual max rating on CF are all > 2900.

So IMHO this team has very high chance of winning WF this year. Is there any reason you put them at 4th? Is it because of performance on OpenCup and Petrozavodsk?

→ Reply
  • »
    »
    8 years ago, # ^ |
      Vote: I like it +48 Vote: I do not like it

    Probabilities are just my intuition.

    CF rating is a good indicator, but CF is an individual contest and uses different format. Previous ICPC and NEERC are important, but they are only two contests. When we have the results of 28(!) ICPC-style contests like SPb ITMO or SPb SU, the prediction can be quite accurate.

    Anyway, if we include NEERC, ITMO vs SU = 17 vs 11, so they are quite close.

    When we don't have much data like Seoul, I mainly depended on ratings.

    → Reply
»
8 years ago, # |
  Vote: I like it +41 Vote: I do not like it

Actually just one month ago there was a pre-final workshop in Moscow, I guess the result will also be helpful: Moscow Pre-final Workshop

→ Reply
  • »
    »
    8 years ago, # ^ |
      Vote: I like it +10 Vote: I do not like it

    Yes, I added and modified my prediction a bit.

    → Reply
»
8 years ago, # |
  Vote: I like it 0 Vote: I do not like it

How do you compute the agg(regations?) of the CF max ratings of team members?

→ Reply
»
8 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Auto comment: topic has been updated by rng_58 (previous revision, new revision, compare).

→ Reply
»
8 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Auto comment: topic has been updated by rng_58 (previous revision, new revision, compare).

→ Reply