What this team's results have earned: every match in the database is replayed in date order, both sides start at 1500 and the winner takes points from the loser. The rating is the one it held after its last rated match, and the rank counts every club ever rated, sides that no longer play included. A measure of strength, not a forecast of the next result. Computed 3 Nov 2026.
| # | Run | From → To | |
|---|---|---|---|
| 1 | 19 | 24 Sep 2006 → 10 Jun 2007 | ▸ matches |
| 2 | 19 | 12 Jun 2005 → 06 Nov 2005 | ▸ matches |
| 3 | 16 | 23 Aug 2014 → 01 Mar 2015 | ▸ matches |
| 4 | 16 | 03 Nov 2002 → 29 Jun 2003 | ▸ matches |
| 5 | 15 | 02 Mar 2014 → 10 Aug 2014 | ▸ matches |
| 6 | 14 | 21 Mar 2008 → 17 Jun 2008 | ▸ matches |
| 7 | 13 | 24 Jun 2004 → 17 Oct 2004 | ▸ matches |
| 8 | 12 | 15 Apr 2006 → 27 Jun 2006 | ▸ matches |
| 9 | 12 | 28 Oct 2004 → 13 May 2005 | ▸ matches |
| 10 | 11 | 27 Apr 2002 → 27 Jul 2002 | ▸ matches |
| Competition | P | D | Draw% | N | Max |
|---|---|---|---|---|---|
| Premier League | 379 | 68 | 17.9% | 5 | 19 |
| Armenian Cup | 49 | 5 | 10.2% | 2 | 4 |
| UEFA Europa League | 6 | 2 | 33.3% | 1 | 2 |
| Year | P | D | Draw% | Max |
|---|---|---|---|---|
| 2015 | 32 | 7 | 21.9% | 7 |
| 2014 | 31 | 1 | 3.2% | 15 |
| 2013 | 60 | 18 | 30.0% | 9 |
| 2012 | 2 | 1 | 50.0% | 1 |
| 2011 | 8 | 0 | 0.0% | 8 |
| 2010 | 34 | 6 | 17.7% | 7 |
| 2009 | 32 | 5 | 15.6% | 8 |
| 2008 | 32 | 8 | 25.0% | 14 |
| 2007 | 32 | 6 | 18.8% | 12 |
| 2006 | 28 | 2 | 7.1% | 12 |
| 2005 | 30 | 3 | 10.0% | 19 |
| 2004 | 31 | 6 | 19.4% | 13 |
| 2003 | 32 | 4 | 12.5% | 12 |
| 2002 | 26 | 4 | 15.4% | 11 |
| 2001 | 24 | 4 | 16.7% | 7 |
Over 434 matches (Apr 2001 – Nov 2015), Ulisses drew 75 where the draw model expected 91.6 from those particular fixtures — about 17 fewer, or 2.0 standard errors, a gap most teams show over a stretch this long.
As unusual as a coin landing the same way 4 times running. A flat league average would have expected 88.7 instead — "more draws than the league" and "more draws than these fixtures deserved" are different claims. These matches were part of the model's own training set, so its chances were fitted already knowing the results.