| Elo Rating calculations for the 2000 NZ Tantrix Championship only | |||||||||
| December 2000: Now allows for the change in calculation of G: G is now simply the number of games played | |||||||||
| Six players had pre-tournament Elo ratings which were official: | |||||||||
| Pekko | Matt Peek | 2023 | |||||||
| mikem | Mike McManaway | 1957 | |||||||
| Justin | Justin Phillips | 1947 | |||||||
| Ridcully | Jamie Sneddon | 1937 | |||||||
| Runes | Phil Sneddon | 1862 | |||||||
| Scarab | Murray Pharaoh | 1748 | |||||||
| Players who previously had unofficial or no ratings | |||||||||
| For such players, the rating for a simple single elimination tournament like this is calculated as follows: | |||||||||
| 1) If such a player played one or more rated players, their ratings are based on those games and their opponents' pre-tournament ratings. | |||||||||
| 2) Their ratings are then used to work out the ratings of the players they beat. | |||||||||
| 3) This process carries on until all ratings have been calculated. | |||||||||
| This means that in most cases we can work back along each eighth of the draw separately. | |||||||||
| In the tables: | |||||||||
| - G is the measure of number of games played as described in section 4 of the rating method. | |||||||||
| - TP % is the weighted (by G for each match) average tournament point percentage | |||||||||
| - Factor is the Elo difference factor for the given TP % | |||||||||
| - Oppnt Elo is the weighted (by G for each match) average opponent Elo rating | |||||||||
| - Tournt G is G for the whole tournament | |||||||||
| - Pre-tnmt Elo is the player's pre-tournament Elo rating | |||||||||
| Player details | Calculation for match v officially rated player | Tournt | |||||||
| Name | Nickname | Lost to | TP % | Factor | Oppnt Elo | Elo | G | ||
| Matt Kearse | MattK | Pekko | 43.8% | -89 | 2023 | 1934 | 8.0 | ||
| Rochelle Duke | Rochelle | MattK | 40.0% | -143 | 1934 | 1791 | 4.0 | ||
| Bevan Chong | Mighty | Pekko | 37.5% | -180 | 2023 | 1843 | 7.0 | ||
| Robin Brown | robin | Mighty | 18.3% | -511 | 1843 | 1332 | 7.0 | ||
| Paul Martinsen | Ponder | robin | 47.5% | -35 | 1332 | 1297 | 4.0 | ||
| Alexei Drummond | venus | Ridcully | 35.0% | -218 | 1937 | 1719 | 5.0 | ||
| Ekarin Teng | nArike | venus | 18.3% | -511 | 1719 | 1208 | 3.0 | ||
| Damon Horrell | Damon | Justin | 45.0% | -71 | 1947 | 1876 | 8.0 | ||
| Jo Simons | SusanDeath | Damon | 41.3% | -125 | 1876 | 1751 | 4.0 | ||
| Chris Peek | Chriz | Runes | 47.5% | -35 | 1862 | 1827 | 8.0 | ||
| Jason Cornish | Jace | Chriz | 37.5% | -180 | 1827 | 1646 | 8.0 | ||
| Alida Remiens | JustAsk | Jace | 36.3% | -199 | 1646 | 1447 | 4.0 | ||
| Mark White | mwhite | Runes | 13.3% | -628 | 1862 | 1234 | 7.0 | ||
| Matthew Boxer | Boxer | mwhite | 33.8% | -237 | 1234 | 996 | 4.0 | ||
| Ross Johnstone | rossj | Scarab | 47.5% | -35 | 1748 | 1713 | 4.0 | ||
| Revel Drummond | Revel | Scarab | 37.5% | -180 | 1748 | 1568 | 8.0 | ||
| John Yamasaki | tyedyed | Revel | 31.3% | -276 | 1568 | 1291 | 4.0 | ||
| Lisa Stowell | windy | mikem | 25.0% | -382 | 1957 | 1575 | 4.0 | ||
| Players who previously had official ratings | |||||||||
| In single elimination tournaments, a player with an official pre-tournament rating is rated as follows: | |||||||||
| - use the normal method for matches played against opponents who had official pre-tournament ratings | |||||||||
| - use the normal method for matches against any other opponents who played more than one match | |||||||||
| - give the player his/her own pre-tournt rating for matches against opponents without official ratings who did not play anyone else | |||||||||
| - take an average weighted by G against each opponent to calculate the overall rating for the tournament | |||||||||
| For an explanation of this method, please see (*) at the end of this page. | |||||||||
| The key to the table can be found above the previous table on this page. | |||||||||
| Player details | Matches v officially rated players only | Pre-tnmt | Tournt | Tournt | |||||
| Name | Nickname | G | TP % | Factor | Oppnt Elo | Elo | Elo | G | Elo |
| Matthew Peek | Pekko | 12.0 | 51.3% | 18 | 1912 | 1930 | 2023 | 16.0 | 1953 |
| Mike McManaway | mikem | 11.0 | 44.7% | -76 | 1457 | 1381 | 1957 | 14.0 | 1504 |
| Justin Phillips | Justin | 3.0 | 31.7% | -270 | 1937 | 1667 | 1947 | 7.0 | 1827 |
| Jamie Sneddon | Ridcully | 11.0 | 40.8% | -132 | 1502 | 1370 | 1937 | 15.0 | 1521 |
| Phil Sneddon | Runes | 8.0 | 57.5% | 107 | 1947 | 2054 | 1862 | 15.0 | 1964 |
| Murray Pharaoh | Scarab | 3.0 | 25.0% | -382 | 1957 | 1575 | 1748 | 11.0 | 1701 |
| (*) | |||||||||
| Single elimination tournaments where less than half the players are officially rated are difficult to rate fairly. | |||||||||
| The first principle is that all games played in a tournament need to be counted in G. | |||||||||
| - this is because G determines whether a player has an official rating or not at any point in the future. | |||||||||
| Using the normal method can cause distortions if a rated player beats an unrated player by a huge margin. | |||||||||
| Using the normal method in this case can actually drag the previously rated player's rating down. | |||||||||
| - this is because the averaging of TP % and Oppnt Elo pulls in different directions | |||||||||
| - if there are no extreme results, these effects cancel out in calculating the rating | |||||||||
| - in an all-play all, extreme results have very little effect in the overall averages | |||||||||
| - but the distortion can be large when: | |||||||||
| --- a) the unrated opponent does not play anyone else | |||||||||
| --- b) the rated player plays only two or three more matches in the tournament | |||||||||
| All this is best explained with an example which can be provided for anyone who is interested. | |||||||||
| Giving a player his/her pre-tournament rating for matches against opponents who do not play anyone else effectively removes this distortion. For players with no extreme results, it will make negligible difference to their ratings. | |||||||||
| An alternative of rating based only on results against rated players can also cause distortions in the overall ratings because G for the tournament may be quite large and if the one or two results against rated players were very lucky or unlucky, they will be given a disproportionate weighting in the player's overall rating. | |||||||||