The early plan was always a Keshik rush.
The problem I usually have with a Keshik rush is the amount of technology required to pull it off. Horseback Riding is expensive, and then you still need Archery, Bronze Working for chopping and slavery, and worker techs for, well, production essentially. The difficulty is that while I'm trying to set up a production heavy empire, the time to research Horseback Riding in particular just drags out and I lose my timeframe for an early rush.
However, in this two-civ setup, one civ can keep a reasonable economy going while the other goes all out for the rush. So while Mongolia sets up for a rush, England can research the key techs. With the slightly faster tech pace as well, I saw the micro for a Keshik rush as being more similar to a standard chariot rush than anything else. For a chariot rush, you build two cities, then chop and whip like mad up to 10-20 odd chariots and send them in before 1000BC. So, that was roughly what I did, except 4 cities (including 2 capitals) because I had two civs, and 3 of them Mongolian because they make the Keshiks.
Mongolia built the settlers (with Imperialistic) while England built the workers. Then Mongolia built military while England built cottages on the flood plains in the capital and kept the economy going.
When we met France/Ethiopia quite early on, I started sizing them up as the target. I kept civil diplomatic relations though, even negotiating a settling border, to keep my options open. But without making any sort of NAP of course. However, later when I met Rome/Netherlands, and discovered their geographic situation, things changed. Basically, they had Rome, and they had literally one team as their neighbour: me. There was no way they weren't going to build Praetorians, and there was no-one else they would want to send them to than me. Especially if I left my back exposed while rushing another neighbour.
This was the first big decision of the game. T. Claudius and I discussed it at great length. Basically it was the proximity of France/Ethiopia against the Praetorian threat from Rome/Netherlands. In the end Rome/Netherlands were the choice, and I think it was the right one.
France/Ethiopia were meanwhile busy taking out the remaining and abandonned half of the rather unfortunately split Maya/France (or should I say Maya ... ocean ... France). I sent them a long message explaining why their turning on me while I'm distracted in Rome/Netherlands would simply bog us all down in war and lead to an easy run-away victory for Lord Parkin, who had taken out the other half of Maya/France and had apparently a continent to himself. This kept them off my tail, despite Rome/Netherlands pleading to come to their aid, although it ended up getting me into some trouble later on.
The broad plan was to rush in with Keshiks, pillage and raze whatever I could, cutting metals if possible (it wasn't in the end). Then, while he's spamming out spears like mad in response, I follow through with a big fat stack of axes (and catapults as it turned out). It played out exactly how I wanted it to in broad terms. I made some horrible mistakes along the way, and Grant played quite well, but the broad plan was good and it worked. The Keshiks razed only two cities at first, but kept him bogged down and distracted. He sent Praetorians in a long march to counter-attack, but by the time they arrived they saw axes and catapults a plenty, and simply turned around. So most of his army never even fought a battle. When some Keshiks circled around and surprised him by razing two more cities he surrendered.
My biggest mistake is worth mentioning, and came right at the start. I captured a worker, but saw there were too many units in the adjacent city for me to capture it, and sent the bulk of the Keshiks elsewhere. This turn, I should have deleted the worker, as it was bound to be recaptured, but I simply didn't think of it at the time. Because I didn't delete the worker, those units got to my launchpad city a turn earlier than they might have, catching it with two less units than it ought to have had defending it. The worst part however was that the gold from capturing this city let him upgrade a warrior to a spearman. It was only this upgrade which prevented me from capturing the Roman capital on the second or third turn of the war! Things could have gone a lot faster for me if I'd deleted that worker.
But as I said we won the war eventually anyway. The offer that Grant made in his surrender was the second big decision of the game.