Got a set of Tenergy 2600 mAh AA NiMH cells a while ago and ran some self-discharge tests.

The three curves on the far right (two overlap pretty closely) are freshly charged. All three packs exhibit pretty much the same 1.8 Ah capacity, about 1.8 Ah or 70% of the advertised value.
I must say that Tenergy cells aren’t exactly Tier 1 quality.
The little stubby red curve on the far left is the as-received capacity. They’re not advertised as ready-to-use and, for sure, they’re not. My analysis of some RTU cells is there.
The two curves on the left of the main group, after six and seven weeks, are down to 1.4 Ah. That’s 78% of the actual fresh-charge capacity, a mere 54% of nominal.
Condensing the endpoints into a table, we have:
Pack | Charge Date | Discharge Date | Storage Time | Capacity |
A | 09/04/09 | 09/04/09 | 0 | 1.82 |
A | 09/04/09 | 09/11/09 | 7 | 1.64 |
B | 09/04/09 | 09/24/09 | 20 | 1.58 |
C | 09/04/09 | 10/25/09 | 51 | 1.41 |
A | 09/11/09 | 10/25/09 | 44 | 1.41 |
B | 09/24/09 | 10/25/09 | 31 | 1.48 |
A picture being worth a kiloword, a quick-and-dirty graph is in order:

Ignoring the first and last points, the slope is more-or-less constant at 6 mAh/day.
Call it 0.3%/day after the first week: store the cells for 50 days and you’re down 15%.
Remember, that’s measured from the actual as-charged and as-measured capacity, not the nominal as-read-about-in-the-spec-sheet capacity.