Statistics

There are some hidden, persistent regularities behind the apparently independent and randomly scattered numbers. Relying on multiple sources, this section provides an insight into the long-term IYPT trends and statistics.


Growth of the IYPT

The table presents the growth of the IYPT in terms of participating nations, teams, grading jurors, and given grades. Click on the headers to have the table sorted by any desired parameter.

YearNTONTPNTGNCONJNstNGNJ-FITNG-FIT⟨SP⟩/6
1988313103--0--00--
198984487--8--211489.04
19906665--27--165017.04
19917976--24--194458.13r
199212121210--52--329657.86
199319191811--84--4816336.72r
199412121211--49--329097.12r
199515151512--57--4010587.16
199613131310--51--349477.07r
199715151511--57--4010587.20
199818181816--66--4812257.28
199919191917--69--5012817.15
200017171716--63--4511696.84
200118181816--93--4817267.28
200220202018--103--5319126.79
200323232322--118--6121906.87
200426262624--133--6924697.12r
20052525252361128--6623767.21
200626262524--128--6623766.89
200722222221--113--5820976.47
200821212121--108--5620056.83
200927272727--138--7225616.26
2010232323236811925146122096.72
2011212121215610819355620056.22
2012282828288014326197426545.59
2013262626267313328296924695.84
2014282828286714425597426735.75
2015272727276813824487225615.65
2016292929297714924997727665.59
Total582620549----2603--145748388--

The columns are:

  1. Year of each IYPT event, starting from the 1st IYPT in 1988;
  2. NTO, officially recognized number of participant teams;
  3. NTP, teams present in the premises of the IYPT. In 1989, 38 Soviet teams from all 15 constituent Soviet republic attended, but only two of them qualified for the International Finals after semi-official selective fights, in which non-Soviet teams participated irregularly and rather as learning observers. In 1991, observing teams of France and Italy attended, but did not compete at all;
  4. NTG, teams obtaining official grades from the jury during the IYPT. In 1988, no rigorous official ranking existed and “[t]he organizing committee set no goal to determine winners”. In 1989, only participants of the International Finals on March 31 were considered participants of the 2nd IYPT. Team of Moldova in IYPT 1993 and team of Nigeria in IYPT 2006 were considered official entrants but did not participate in Physics Fights and thus obtained zero scores in the ranking;
  5. NCO, number of different countries represented by teams. In 1988, this number was considered 3 despite presence of 14 constituent Soviet republics (all except Belarus). In 1989, two Soviet-based teams (from Russia and Ukraine) were both considered from the USSR. In 1990, team of Rīga is counted as Soviet-based despite its complex political status. In 1991, the Moscow-based and the Soviet-combined teams are considered both from the USSR;
  6. NJ, number of jurors giving grades used in the official ranking. This number is known with an absolute certainty only for 2005 and 2010—2016;
  7. Nst, number of officially graded IYPT stages. This number is set to 0 in 1988 and to 8 in 1989, albeit there were up to 97 friendly, training selective stages in 1988 and up to 140 in 1989. It is thus known with a very high degree of certainty for each IYPT;
  8. NG, number of official grades given to teams and used to produce official ranking. This number is known with an absolute certainty only for 2010—2016;
  9. NJ-FIT, most likely number of jurors giving official grades. This number is obtained by fitting NJ to NTG in the known years and using the best-fit parameters to calculate the expected number in the unknown years. Note that difference between NJ-FIT and NJ in the known years never exceeds 7 persons and on average equals 4 persons (6% average error);
  10. NG-FIT, most likely number of grades given to the teams to produce official ranking. This number is obtained by fitting NG to Nst in the known years and using the best-fit parameters to calculate the expected number in the unknown years. Note that difference between NG-FIT and NG in the known years never exceeds 360 grades and on average equals 181 grades (7% average error);
  11. ⟨SP⟩/6, average Sum of Points (in all IYPT rounds, including Semi-Finals and Finals) divided with 6. It is an indication of the average or the most popular grade at the IYPT, although not equal to ⟨G⟩ since reports, oppositions and reviews contribute to SP with different weights.



Forecasts

Since 1988, NCO growed linearly with time. By fitting this dependence, it is easy to forecast the expected number of teams in the future IYPTs. In the past IYPTs, the values of the fit missed the actual values by 1.8 teams on average.

YearNCNJNgr
201731   8210
201832   8410
201933   8611
202033   8911
202134   9111
202235   9311


Involvement of countries

The table presents a list of countries ordered by the first year of participation in the IYPT. In case of equal years, they are ordered alphabetically. Also presented are number of teams from the country in the Finals, the last year in the Finals, total number of teams from the country, and the last year a team competed.

45 different countries have competed in the IYPT, but 14 of them did not compete in the last two years. No longer existent Soviet Union and Czechoslovakia are listed separately from Russia and Czech Republic, respectively. Click on the headers to have the table sorted by any desired parameter.

No.CountryJoinIn FinalsLastAll teamsLast
1. Bulgaria19882   198919 2016
2. Czechoslovakia19883   19925   1992
3. USSR19887   199135 1991
4. Germany198915 201625 2016
5. Hungary19895   199725 2016
6. Netherlands19893   199219 2012
7. Poland19899   201534 2016
8. UK19910 --12 2016
9. Belarus19924   200524 2016
10. Georgia19924   199921 2016
11. Kazakhstan19920 --1   1992
12. Moldova19920 --2   1993
13. Russia19923   200040 2016
14. Ukraine19921   199327 2016
15. Czech Republic19935   199825 2016
16. Slovakia19933   201424 2016
17. Uzbekistan19930 --7   1999
18. Sweden19940 --21 2016
19. Finland19950 --10 2009
20. Armenia19960 --1   1996
21. Australia19982   200719 2016
22. Austria19984   201120 2016
23. Mexico19980 --7   2006
24. United States19991   20058   2016
25. Croatia20022   200813 2016
26. South Korea20028   201316 2016
27. Switzerland20022   201616 2016
28. Indonesia20030 --8   2013
29. New Zealand20034   201013 2016
30. Brazil20040 --10 2016
31. Cyprus20040 --2   2007
32. Kenya20040 --9   2015
33. Nigeria20060 --8   2015
34. China20082   20159   2016
35. Iran20081   20128   2016
36. France20090 --4   2013
37. Singapore20096   20168   2016
38. Thailand20090 --7   2016
39. Chinese Taipei20101   20167   2016
40. Slovenia20120 --3   2014
41. Romania20130 --4   2016
42. Macao20140 --3   2016
43. UAE20140 --1   2014
44. Canada20160 --1   2016
45. Pakistan20160 --1   2016


Presented problems

Some of the IYPT problems are more popular among entrants and are presented more often than other problems. The table traces how many times each problem has been presented in the IYPT. Note that within each IYPT the total number of presentations strictly equals Nst.

As the IYPTs are not equal in size, it may be meaningul to normalize the data with Nst/17 to obtain relative popularity of problems and compare various IYPTs. In the available dataset of 15 IYPTs, only two problems have exceeded average popularity by a factor of 2 or more: No. 8 “Magnetic train” in 2016 (relative popularity of 2.40) and No. 13 “Air lens” in 1995 (relative popularity of 2.39.)

Year0102030405060708091011121314151617Σ
1988----------------------------------0
1989----------------------------------8
1990----------------------------------27
1991----------------------------------24
1992----------------------------------52
1993----------------------------------84
1994----------------------------------49
19952123460515618331657
19962463426423331134051
19972134332636334154457
19987641235721737321566
1999----------------------------------69
2000----------------------------------63
20015769575245563927693
2002----------------------------------103
2003----------------------------------118
2004----------------------------------133
2005----------------------------------128
2006857861010769116528128128
20071166687793865461056113
2008----------------------------------108
200910979310771010935118146138
201068987999746489556119
2011669410359465778658108
20121691171210118789578780143
20137913711128751251084564133
20147106611815106759897155144
2015528911156512879149387138
2016997916613214710974756149


Rejected problems

Some of the IYPT problems are hard to solve and a challenge is likely to be rejected. Some other problems are easier or more attractive and can be accepted or selected by many teams without a single reject. The table presents the number of times that each problem has been rejected when challenged at the IYPT. Note that within each IYPT the total number of rejected challenges is not constrained and is an interesting parameter on its own.

Year0102030405060708091011121314151617Σ
1988------------------------------------
1989------------------------------------
1990------------------------------------
1991------------------------------------
1992------------------------------------
1993------------------------------------
1994------------------------------------
19957113320130070113134
19965103003102130113024
19971112103023311101223
1998------------------------------------
1999------------------------------------
2000------------------------------------
20011231312551324063244
2002------------------------------------
2003------------------------------------
2004------------------------------------
2005------------------------------------
20062522449483342233262
20070143353344325346154
2008------------------------------------
200922738244312134252569
20102222130137371043748
20111604452371100035345
2012252625427951142641086
20134404422261653453055
20147367434451972441778
20153354039225532184463
201602443621711413505967


Ranking dynamics

Yearσ(Δr12)σ(Δr23)σ(Δr34)σ(Δr45)Best-fitPF 0.5σ(Δr1F)
1988----nonenone------
1989----nonenone------
19901.151.000.00none2.442-0.575*x3.381.15
1991----nonenone------
1992------none------
1993--------------
19942.001.831.48none2.550-0.260*x7.881.88
19952.941.900.67none5.242-1.135*x4.183.76
19961.811.88--none------
19972.371.631.83none2.753-0.270*x8.344.32
19983.231.642.01none4.123-0.610*x5.942.98
19992.963.241.33none4.955-0.815*x5.475.13
20002.452.151.33none3.657-0.560*x5.642.29
20013.312.081.941.274.341-0.626*x6.144.83
20022.891.601.490.843.896-0.626*x5.423.52
20033.692.862.111.625.006-0.696*x6.475.44
2004--------------
20052.882.912.021.394.176-0.536*x6.864.77
20064.152.351.771.525.412-0.847*x5.804.30
2007--------------
20082.682.051.311.543.351-0.416*x6.853.60
20092.352.431.441.363.281-0.396*x7.022.48
20104.102.881.661.175.956-1.001*x5.454.03
20113.332.451.781.164.693-0.718*x5.845.51
20123.092.262.211.623.856-0.446*x7.523.72
20132.592.121.361.393.391-0.436*x6.633.01
20143.231.711.471.254.078-0.618*x5.793.79
20153.603.101.421.635.094-0.759*x6.055.77
20163.102.021.790.844.391-0.701*x5.554.21


Juror grades

The diagram shows the distribution parameters (population mean and standard deviation) for grades of individual jurors at the 23rd IYPT (2010.) The grades are averaged across all stages, performances, and PFs. The data clearly reveals that the individual grading standards vary among jurors, and result in a marked spectrum of average grades in the range from ca. 6 to ca. 8. The relative standard deviations, not shown here, vary in the range from ca. 0.1 to ca. 0.3. The marks above 8—9 and below 5—6 are statistically rare events.


Theoretical problems

It was quite typical at the early IYPTs to dig into problems about gold cubic planents, Grand Unification, or Asrtid Lindren’s Karlsson-on-the-Roof, where only a theoretical solution was expected. In some cases, the problems were formulated in a way that only limited empirical verification was possible as such. The annual number of theoretical problems, however, has been dropping sharply from ca. 10 in late 1980s, to only a few by late 1990s. Starting from 2003, each of the 17 problems was permitting a straightforward experimental investigation. (But not every solution, naturally, was using this opportunity.) The half-life of the theoretical IYPT problems can be estimated as 5 years, as the graph shows.


Stage regulations

In 1988, the IYPT was starting with only 5 min allocated for Report, 3 min allocated for Opposition, and 2 min allocated for Review. The duration of each stage performance, especially for the Report, was increasing stepwise until mid-1990s. It was, however, only in 2009 when the Reviewers were allowed to take floor for 4 min.


Growth over years

The graph depicts the growth of the IYPT in terms of participating nations and teams. Cumulant nations, or the total number of countries that have taken part at a IYPT, grows steadily with time at a mean rate of 1.5 new countries per year. The instant fluctuations of this rate are illustrated by the data for New nations, joining the particular IYPT for the first time ever; the decceleration trend (from expected 2.2 new countries in 1990 to expected 0.9 in 2010) is disputable given the limited statistics and is influenced by many new-born Eastern European nations in 1992—1993. The nations present at a particular IYPT increasingly lag behind the cumulant nations (with their mean growth rate of 1.0 per year.) The teams present at a particular IYPT usually outnumbered the present nations, until this was formally banned in 2008; the mean difference between them, as for 1989—2007, was 2.1, reminiscent of the policy that the host nation and Russia could be represented by two teams, and further teams might be permitted by LOCs.