"Bouncing" is a tactic wherein the bulk of a siege force moves during a month. A cover force holds the siege during the time in which the main siege force is elsewhere.
Bouncing a siege makes it possible to advance two sieges using two cover forces, and one additional force large enough to siege. A second use for bouncing is to reduce attrition in a siege which is taking place in a province which cannot support a siege force without attrition.
Siege Rules Exploited in Bouncing
According to the rules for sieges, a siege is only advanced on one day per month. As long at least a cover force is left in the province the rest of the time, siege advancement will take place. At the time of siege advancement, if there not enough force present to siege, then siege advancement will be stopped. If enough force is present, then the siege advancement timer will be set (for one month). After the timer is set, the next siege advancement will take place so long as at least a cover force is left in the province at all times; a full siege force is not needed.
Thus, we can see that the full siege force is really only necessary in a province 1 day per month: to set the next siege advancement timer. The rest of the time, only a cover force (roughly 20%) is needed. The bulk of the force (80%) is free to move around the rest of the month.
Typical Instances of Bouncing
Here are some typical cases where I use bouncing.
Advancing Two Sieges
Consider a situation where there are two plains provinces A and B. Two cover forces are present, one per province, each with 1200 men. The bouncing force, 4000 cavalry, ride back and force between the two provinces. Each province ridden takes 10 days; the round trip takes 20. Now, assume the siege advancement day in province A is the 1st, and in province B, it is the 11th. We can then do as follows:
- on the first of each month, province A siege advances. Pause.
- Order bouncing force from A to B.
- On the 11th, force is present in B. Province B siege advances.
- Order bouncing force from B to A.
Both sieges are advancing as fast as a siege can. Yet a total of only 6400 troops is covering both sieges.
You can see that by bouncing we've increased our siege efficiency. If we define the use of 5000 troops for one month as a standard siege unit (SSU), then we can see that a typical siege (of a level 1 fort) will achieve roughly 1 siege-advancement per SSU. By bouncing, we've increased that to 2 siege-advancements per 1.2 SSU; that is, our siege rate has increased by 66%. Nice!
You can use cavalry to advance two sieges in adjacent provinces, so long as the land movement times for the two sum up to no more than 29 days. If it is longer, of course you can still bounce, but it will cost you in movement attrition. (If the total round trip time is more than 49 days, bouncing no longer speeds up sieging.) Generally I recommend against this, unless you've gotten to a point in your game where money has ceased to limit your ability to recruit. Of course, in that case you probably don't need to bounce; just siege every province. It would be too much micromanagement in any case.
Now consider a case similar to that above, except using infantry as the bouncing force. In that case, you can only move once per month. (The assumption here is that movement attrition is to be avoided; obviously we can move twice per month if not.) So we have the following loop:
- On the first of even months: Pause. Order bouncing force from A to B.
- On the 15th, bouncing force gets to B. Province B siege resumes. Province A siege advances (it was restarted a month back); however, it now gets suspended due to only having a cover force there.
- On the first of odd months: Pause. Order bouncing force from B to A.
- On the 15th, bouncing force gets to A. Province A siege resumes. Province B siege advances (from last month). It gets suspended due to only having a cover force there.
In this case, both sieges are not advancing as fast as a siege can. Rather each is advancing once every other month, i.e., at half speed. If we compute siege efficiency, it is 0.86 siege advancements per SSU; less than 1 SA/SSU.
So why would you do this? One reason is that on siege completion, you don't need a unit present to re-siege; in that case, having the full siege force present is wasting time. If the odds are greater than 16% that we capture a fort, then bouncing can speed up sieging.
A second reason is that covering up AI countries is a vital tactic, due to the their ability to get income from uncontrolled (but owned) provinces, to mint endlessly without destroying their economy, and thus to recruit armies continuously. In general the only way to get 100% warscore reasonably quickly is to destroy the enemy army, then cover every province he owns. In this sort of situation, a cover force is required on every province, anyway. Thus we should compute siege efficiency without counting the cover forces; and here we find siege efficiency is just as good when bouncing infantry as it is without. And it will be better (see the last paragraph) once we get to a net modifier of +4 in the siege, or better. (This happens with a normal unaugmented siege when the siege state drops to -4.)
Bouncing to Avoid Attrition
A bouncing cavalry force can lower or eliminate attrition in a siege. Consider a case where you are sieging a city with base tax value of 2. In this case, you can supply just 8 units in the province without attrition. (More if you have a leader with a good maneuver value, but this is usually not the case.) What if there is a small fort in that city? Well, one thing you could do is stick in slightly more than 10000 men, and just take the 3% attrition every month. Losing about 300 men per month, you just feed in another 1000 men every three months until the siege completes. This can get expensive!
But what if there is an adjacent province with some extra support capacity? In that case, we can leave a cover force of 2000 men (or 2400) in the siege province. Then we take our cavalry army and put it in the adjacent province: it moves in, restarts the siege, then moves back out, once per month. Because it ends the month in the other province, and by assumption there is enough support capacity there for it, no attrition happens to either force. Yet the siege advances normally, once per month.
Bouncing to avoid attrition is an important tactic later in the game (where fortress levels have become high, requiring very large siege armies). It is also important early in the game for fighting wars in winter.
Bouncing with Slow Forces to Avoid Attrition
Another case that comes up at times is a situation of potentially high attrition, where the forces you have available are too slow to get in and out of a siege province in a month. This can happen anywhere in thick mountains, which even using a cavalry bouncing force, would require a 35 day round trip. Or it might happen across a river, if that's the only access you have to a province. But it also happens in a few places with slow movement, most particularly Africa. And it is necessarily the case for aboriginal American countries before they can build cavalry.
In cases like these, there is no way to siege without taking attrition. However, often it is possible to lower the attrition your forces take, to just movement attrition, by bouncing. Typically you have one cover force, and two bouncing forces. Each bouncing force is ordered to move to the siege province in alternating months, calculating movement time to arrive on the first of the (next) month. Once there, it immediately moves back out. So long as the combined move in and out is 59 days or less, this setup can keep the siege force from being present in the siege province at any end-of-month.
Use of Artillery when Bouncing
Of course artillery are very slow; so it is better to let the artillery be the cover force unless there is some attrition for the cover force as may be the case for a winter province. But if you want more artillery than can be supported it is useful to realize that a unit of 10 artillery is small, but is the same as 1000 other soldiers. Because artillery are small in numbers they can avoid loss when there is a small amount of attrition, as in the case of movement attrition.
Units of 10 artillery need repair by splitting, merging, or reorganizing before 10 months passes in the case of 1% movement attrition or more frequently in the case of greater amounts of attrition. Multiple units consisting of only 10 artillery can do the bouncing so long as they arrive early in the month and can flee to a plains, desert, or forest province without crossing a river or strait. Or multiple units consisting of only 10 artillery can fill up to slightly more than the supply level which may be enough for a minimum cover force or for better siege performance, even if other troops are required to bounce in and out.
In either case, only the artillery present and stationed on a siege the day that siege advancement occurs is used to compute the siege modifier given for artillery. In the case of bouncing cavalry or infantry, missing a departure date by one day is usually not so big a deal; it just means that siege advancement will be one day later from then on. But if artillery miss the departure day, the siege modifier will not be used for the siege advancement for one month.