Tsong 2005

Performance0-Rank  0-Score1-Rank  1-Score2-Rank  2-Score3-Rank  3-Score3R-Rank  3R-Score4-Rank  4-Score  NED
Ashkenazy 1981   18  0.3915  0.0110  0.1314  0.2825  0.0717  0.14
Biret 1990   13  0.4531  0.0023  0.1321  0.1327  0.0824  0.10
Blet 2003   7  0.5216  0.0116  0.1513  0.2813  0.238  0.25
Brailoswky 1960   22  0.3620  0.0024  0.1320  0.1326  0.1020  0.11
Chiu 1999   17  0.398  0.036  0.186  0.4010  0.442  0.42
Clidat 1994   20  0.3817  0.0019  0.1123  0.1122  0.0922  0.10
Cortot 1951   31  0.1725  0.0031  0.0532  0.0532  0.0530  0.05
Ferenczy 1956   30  0.1933  0.0033  0.0333  0.0333  0.0333  0.03
Fiorentino 1962   12  0.477  0.0313  0.1112  0.281  0.544  0.39
Fliere 1977   14  0.445  0.064  0.158  0.3819  0.1111  0.20
Francois 1956   2  0.601  0.361  0.351  0.703  0.611  0.65
Hatto 1997   5  0.5329  0.0014  0.147  0.4025  0.1010  0.20
Indjic 2001   4  0.5410  0.0212  0.143  0.4621  0.119  0.22
Jonas 1947   27  0.2421  0.0025  0.1222  0.1232  0.0527  0.08
Luisada 1990   15  0.4430  0.0021  0.0826  0.0825  0.0926  0.08
Lushtak 2004   10  0.5111  0.0215  0.1215  0.2725  0.0914  0.16
Magaloff 1977   24  0.3126  0.0027  0.0727  0.0728  0.0729  0.07
Michelangeli 1971   29  0.2128  0.0029  0.0629  0.0632  0.0531  0.05
Milkina 1970   16  0.4412  0.0111  0.1216  0.2530  0.0619  0.12
Mohovich 1999   3  0.556  0.038  0.1311  0.3128  0.0715  0.15
Olejniczac 1990   21  0.3732  0.0022  0.1024  0.1026  0.0825  0.09
Poblocka 1999   9  0.519  0.0218  0.1517  0.1532  0.0721  0.10
Rangell 2001   23  0.334  0.075  0.135  0.419  0.365  0.38
Rubinstein 1939   26  0.2718  0.009  0.1210  0.3420  0.1013  0.18
Rubinstein 1952   8  0.5214  0.013  0.152  0.5821  0.127  0.26
Rubinstein 1966   11  0.4927  0.0017  0.1418  0.1425  0.1218  0.13
Shebanova 2002   6  0.533  0.097  0.139  0.379  0.413  0.39
Smith 1975   19  0.3919  0.0020  0.1319  0.1327  0.0823  0.10
Sofronitsky 1960   25  0.2824  0.0028  0.1025  0.1029  0.0628  0.08
Tsong 1993   1  0.612  0.202  0.214  0.4316  0.246  0.32
Tsong 2005   target  targettarget  targettarget  targettarget  targettarget  targettarget  target
Uninsky 1971   33  0.0634  0.0034  0.0334  0.0333  0.0434  0.03
Yaroshinsky 2005   28  0.2123  0.0032  0.0530  0.0532  0.0432  0.04
Random1   32  0.0813  0.0126  0.0628  0.063  0.5612  0.18
Random2   34  -0.0222  0.0030  0.0531  0.059  0.3716  0.14

Note: To load data table give above into Excel, copy and paste the data into a text editor (such as WordPad) first, then copy the text in the editor and past into Excel. You should remove the "target" line from the data before pasting into Excel so that plotting graphs of the data is done properly.

Column descriptions

  • Performance:
  • 0-Rank/0-Score: 0-Score is equivalent to Pearson correlation of the entire data sequence between the reference performance and a test performance. 0-Rank is the sorting order of the 0-scores (highest score has a rank of 1).
  • 1-Rank/1-Score: 1-Score is the area fraction covered by a particular performance in the scape plot (see image above). These values should not be taken literally, since they are sensitive to the Hatto Effect.
  • 2-Rank/2-Score: 2-Score values are equivalent to 1-Score values with all higher-ranking performances removed before the calculation of the area of coverage in the scape is calculated. Improvment over the 1-Rank scores, but still somewhat sensitive to the Hatto Effect.
  • 3-Rank/3-Score: Similar to 2-Rank calculations. The bottom 1/2 of the 2-rank performances are kept constant as a noise floor for the similarity measurement. Then one-by-one the top 1/2 of the 2-rank performances are superimposed with the noise-floor performances, and a 3-score is measured as the area covered in the scape. This measure is not sentisive to the Hatto Effect.
  • 3R-Rank/3R-Score: Reverse 3-rank/3-scores. 3-rankings and scores are not symmetric (A->B values are different from B->A values). So this column represents similarity measures in the opposite direction.
  • 4-Rank/4-Score: The geometric mean between 3-scores and 3R-scores. This column gives the best overall similarity ranking between the various performances (see color codes below).
  • NED: Noise Equivalient Distance (not yet implemented)

Color codes for 3-rank listings:

  • red = strongly similar performance to target
  • orange = moderately similar performance
  • yellow = weakly similar performance
  • green = marginally similar/dissimilar performance
  • white = dissimilar to target
  • blue = false positive (has high 3-rank score but low 3R-rank score)

3-rank/scores are not symmetric, so the 3R-rank/score columns give the 3-rank/scores going in the opposite direction. More matches in the 3-rank column than in the 3R-rank column indicates an individualistic performance, while more matches in the 3R-rank column indicates a mainstream performance.

If a 3-rank and a 3R-rank are both marked as similar to each other, then there is a possible direct relation between the performances. If one is similar to the other but not in the reverse direction, then the similarity is more likely to be by chance (performers randomly chose a similar interpretation).