Quote Originally Posted by Xdeathfire
The unit pool however makes it so that retraining and training troops are about equal. Many times when you are not allowed to train more troops due to the pool, you could actually retrain a unit a half strength. I would rather use a 3 unit pool to retrain 6 half strength units to retain their experience then to train 3 new fresh troops with little to no experience
I think this is one of the better points I've seen so far in this thread. The M2TW unit pool is tracked at least in tenth points (probably finer than that) as opposed to full units, so retraining part of a unit will in fact deduct the correct percent of a full unit from the correct unit pool - i.e. if you have 1.5 mailed knights and retrain a 1/2 strength mailed knight unit, you are left with 1.0 mailed knights in the recruitment pool. So the retraining is not less efficient inherently, it uses up a percent of a point equal to the percent of the retrained unit that was missing, just as a full new recruitment uses a full 1.0 point of the pool.

There is some concern, though, about having to ship troops back to the place they were produced - there's no question that it takes a long time to get this done. Is it worth it? I'm not sure yet, but perhaps. One thing I would put forward, though, is this: it's quite beneficial to avoid higher level troops than you will typically find enemy cities/castles able to produce. The reason is that using such troops essentially requires they be sent home for retraining or swapped out/merged. In my experience it's usually better to stick with troops you'll find commonly available, and then retrain them in the border settlements you capture. This can be accomplished far more quickly than sending them home for retraining, and simultaneously gives all the experience benefits of doing so, just without the long wait. The best part is that the high experience units you'll soon have will quickly equal and exceed their higher tech counterparts in most cases, especially since you can take the time you would've needed to build to those high tech units and instead invest it in blacksmith buildings to upgrade the lower-tech units you intend to use consistently. So it seems to make sense logistically to use more widely available units and be able to retrain them at the front than to employ any of the various other methods discussed here so far. Of course there are limits, but in a lot of cases it seems prudent to sacrifice some stats in the name of wider retraining availability and thus less logistical issues overall.