I've seen efficiency ratings from 70-80% for Ni-Fe from good sources (I have a trimetric 2020 meter and I plan to test for this). What I find interesting in relation to efficiency and hope to find out properly is that the 20 hour rate for a 100Ah is, ummm 100Ah, but the 100 hour rate is 122Ah. So, something is a bit strange in relation to efficiency. So, far my best guess is that the thing about these batteries is not so much the efficiency but the giving up of electrons, or discharge rate. They really don't like big amounts of current being pulled out of them at any one time and this is what these batteries need to be sized for.
Speaking of efficiency:
Sealed Pb-A at 95% and wet cell Pb-A are around 90%. There are differing opinions but sealed Pb-A have a cycle life of between 200-500 (higher from specialty manufacturers) at reasonable DOD. Wet cell varies greatly, from 500-1300ish (again, higher from specialty manufacturers) from what I can gather. These are at %50 or less.
Li Ion batteries that are commonly in use in cellphones and laptops (generally referred to as Lipo, or Lithium polymer), are only rated for 500 cycles max, require very careful balancing and charging voltages and prone to failure through runaway (uncontrolled) chemical reactions (fire!). Obviously these aren't to be used in RE systems and I'm only mentioning it here to avoid confusion as that's what most people think when they think Li Ion.
However, of the technologies mentioned in the wiki article, the technology that's here--now--are LiFePo4 (Lithium Iron Phosphate) batteries. These are are different story. Cells based on this technology have a cycle life at least 1000-2000 and more at lower depths of discharge. However, this technology is new and the various manufacturers of this tech have varying degrees of quality. From what I can gather so far the best manufacturer of this tech at the moment is a123 systems, a US based company made up of graduates from MIT. There are other specialist manufacturers of this type of battery, LionEv makes 100Ah and 200Ah cells and they cost (at the time of writing) $US302.95 and $US519.95 respectively per cell (complete with balancer and wiring gear); each cell is rated at 3.2v, so you'll need 4 for a 12v system. Obviously, this is doable, and are ripe for using in RE systems. There are also some of this type of cell being hand-made to order (duct tape batteries on ebay) and going into electric bikes. They are cells made in China and look interesting. In my opinion the people who are doing the most interesting investigating on new (and old) battery technologies are the EV (electric vehicle) guys (and they do seem to be guys doing it). With this community there is plenty of testing and sharing of information on the web.
There are also super-capacitors with nano tech and these are promising for storage. But we are talking about what's available 'now'--right? Personally, I'd hate to see what happens to a super, or ultra cap if a fissure happens in the dielectric layer. I suppose we'll soon find out as the new Honda FCX uses these for temporary storage for power from it's fuel cell.
I've been researching battery technology for the last 9 months for use in RE systems. It just seemed like the weakest link in RE and worth looking into. In my mind, the best system would be no batteries at all, a kind of Internet-of-Power generation where people could generate electricity like people generate data. I suppose we kind-of have this already, the main thing hampering it are the power generation companies and legislation (in most parts of the world). As they say, the devil is in the details, and with grid tie that is definitely true. So, we are left with storing our own electricity chemically. NiFe seems as good, if not better, than most solutions out there to me. It's the only battery people talk of handing down to their children (or grandchildren) so I'm willing to give them a go.