Breaking Off From Close Combat -- Take #2

[pozanias]

Originally Posted by david kuijt

You're reading an awful lot into the description of the battle, Kevin. What evidence do you have for the existence of those orders? Basing it upon the results of the battle is far too speculative for my taste -- recoils alone will create the cupped arc of Cannae, aided by deploying recessed in a cup in the first place.

I seriously doubt that any of the commentators interviewed Hannibal himself or his lower-level officers that would have passed on those orders. In fact, most of the information we have on the battle comes from Polybius, who was born 15 years after the battle, and so would have been interviewing Roman veterans 40 years after the battle or more -- when very few of them would have been alive. And it is quite unlikely that he ever met or interviewed any Carthaginian (Gallic/Spanish) veterans of the battle.

[content removed]


Cannae works great with Roman Blade combat factors and the various Carthaginian foot fighting at +4 (whether spear, fast blade or whatever). If the Roman commander keeps pressing forward into the gap, trying to break the center, normal combat results will keep pushing the Carthaginians back.

Although I agree with most of what you said, I wouldn't be so dismissive of Polybius and other sources. They're the best we've got. So, I do think its possible that Hannibal did give the order to give ground. However I also think its at least as likely that he didn't give the order, but did anticipate that result (and even encourage that result by his deployment). In other words, he figured his Gauls wouldn't be able to stand up to the Romans and predicted they would fall back. But we'll never know.

The problem is that 'giving the order' and 'anticipating the result' are two very different things. Perhaps we can be satisfied that 'anticipating the result' is already accounted for in DBA. A player can look at combat factors and predict outcomes. And the most likely result of Bd vs. Raider or Spear (which is probably how Hannibal's Gauls should be classified) is a recoil of the Gauls.

[david kuijt]

Originally Posted by pozanias

Although I agree with most of what you said, I wouldn't be so dismissive of Polybius and other sources. They're the best we've got.

Wait, Mark -- I wasn't dismissive of Polybius. I was pointing out that Polybius was writing 50 years later. That isn't dismissive -- that's the truth. It is crucial to use your sources. But you also need to use them critically -- so when you read Polybius, you need to know when he was writing, and for what audience, and what the historical and literary traditions were at that time.

Originally Posted by pozanias

So, I do think its possible that Hannibal did give the order to give ground. However I also think its at least as likely that he didn't give the order, but did anticipate that result (and even encourage that result by his deployment). In other words, he figured his Gauls wouldn't be able to stand up to the Romans and predicted they would fall back. But we'll never know.

Exactly. That was the point I was trying to make -- that Kevin was stating his case too strongly. As you say, it is possible that Hannibal did order some sort of "withdraw all the way across the front" order -- but it is at least as likely (more likely) that he understood how the battle would go and predicted the results. Predicting results of local engagements is how Great Generals become Great Generals. Hannibal knew his troops, and he knew the Romans, so in DBA terms he knew the combat factors and he knew that his guys were going to be pushed back.

Originally Posted by pozanias

The problem is that 'giving the order' and 'anticipating the result' are two very different things. Perhaps we can be satisfied that 'anticipating the result' is already accounted for in DBA. A player can look at combat factors and predict outcomes. And the most likely result of Bd vs. Raider or Spear (which is probably how Hannibal's Gauls should be classified) is a recoil of the Gauls.

Right.

[broadsword]

Again, I think DBA is rather bad at recreating history, because it replaces the "subjective belief" of generals, with a rather unstable stochastic process (a local martingale, actually). What this implies is that in a one on one Bd vs Sp (or maybe 2 on 2) there is a very reasonable bet that the Bd pushes back the Sp. What I've found though is once you throw in multiple independent trials, you lose some of the global correlation between individual elements. That I think is what Kevin is getting at. When I play DBA, as the Carthos the first thing that'll happen is that my Sp will double their first two Bds!

It's almost like you need some kind of joint distribution that has a variable correlation built in, so that you increase the chances that the entire line recoils repeatedly. I submit that using DBA, Cannae isn't all that likely. Some might argue that Cannae was extremely unlikely, but that would make Hannibal extremely lucky, again and again and again. Statisticians prefer not to go down that path, as Hannibal got many, many things right, not just the big battles - he was very good at extracting information from an incomplete "market".

So how best to simulate this? I suggest you could simulate this kind of result by either having recoils more likely when a neighbour has recoiled, but not kills more likely (this is a battle-specific thing) or by having some battle specific modifiers, such as African Sp are at '-1 CF' except if that would cause a kill ...

I just think a global recoil as a combat move might introduce all kinds of other gamey things that might not interact with some other things in 2.2+. Get testing!

[Kevin Boylan]

Originally Posted by pozanias

Although I agree with most of what you said, I wouldn't be so dismissive of Polybius and other sources. They're the best we've got. So, I do think its possible that Hannibal did give the order to give ground. However I also think its at least as likely that he didn't give the order, but did anticipate that result (and even encourage that result by his deployment). In other words, he figured his Gauls wouldn't be able to stand up to the Romans and predicted they would fall back. But we'll never know.

The problem is that 'giving the order' and 'anticipating the result' are two very different things. Perhaps we can be satisfied that 'anticipating the result' is already accounted for in DBA. A player can look at combat factors and predict outcomes. And the most likely result of Bd vs. Raider or Spear (which is probably how Hannibal's Gauls should be classified) is a recoil of the Gauls.

I'm not averse to considering Cannae a debateable case, although the consensus of professional historical scholarship on this topic definitely favors the ordered retreat theory. Yet, other examples I listed are based on historical sources that irrefutably describe -- and sometimes even quote -- commanders ordering troops to retreat from combat with faster enemies. At Bannockburn, for example, Robert the Bruce's Islemen allies (Blade) left their prepared defenses to attack English knights that had charged them but been recoiled. The "anticipating the result" approach signally fails in this case because the result that the Bruce clearly anticipated was that the Islemen would be "quick-killed" if they stayed where they were. So, instead he did something which the current rules prohibit -- he ordered the Islemen to retreat back to safety by breaking off the combat with their faster foes.

[winterbadger]

Originally Posted by broadsword

It's almost like you need some kind of joint distribution that has a variable correlation built in, so that you increase the chances that the entire line recoils repeatedly.

So how best to simulate this? I suggest you could simulate this kind of result by either having recoils more likely when a neighbour has recoiled

But don't we have that already, in that when one is fighting in a line, the recoil of one element causes the odds to increase that adjoining elements will be recoiled or destroyed, as they are now subject to -1 overlap modifiers that they were not liable to before?

[david kuijt]

Originally Posted by broadsword

Again, I think DBA is rather bad at recreating history, because it replaces the "subjective belief" of generals, with a rather unstable stochastic process (a local martingale, actually). What this implies is that in a one on one Bd vs Sp (or maybe 2 on 2) there is a very reasonable bet that the Bd pushes back the Sp. What I've found though is once you throw in multiple independent trials, you lose some of the global correlation between individual elements. That I think is what Kevin is getting at. When I play DBA, as the Carthos the first thing that'll happen is that my Sp will double their first two Bds!

It's almost like you need some kind of joint distribution that has a variable correlation built in, so that you increase the chances that the entire line recoils repeatedly.

Nah. You just need a few reserve elements (on both sides). If you play Cannae with no reserves, then yes, statistical outliers will usually dominate the results overall.

But that's true about DBA any time. If you want to be vulnerable to the whims of chance, play with no reserves.

Originally Posted by broadsword

I just think a global recoil as a combat move might introduce all kinds of other gamey things

Exactly. Especially if it was a choice by the general.

[Redwilde]

Originally Posted by david kuijt

Especially if it was a choice by the general.

I like that overall DBA does not allow the general to have choice over everything, this seems more reflective of typical history. There are other rules that do allow the general to have choice over everything, and I like some of those too. (Rules According to Ral )

For trying to recreate the exact flavour of any particular battle, DBA may not be your best choice. I'm good with that.

For all of these examples, you can still get the withdrawal in DBA by waiting for recoils to happen to you, and then pulling back. Won't always happen in a tidy and safe way. Probably wasn't a gauranteed safe maneuver historically either. The examples are case studies when the probabilities rolled good that day.

[Lobotomy]

Originally Posted by Kevin Boylan

Larry,

I'm not certain exactly what you are arguing in your first paragraph. Cannae is a clear and totally unambiguous case of troops, in contact with an enemy that was just as fast as they were, who were ordered to retreat -- and did so. If DBA cannot perfectly model HOW they did it, that's hardly a surprise as it's very simple and generic rules set. DBA does not concern itself with process, only with results. As I said, there might be other ways of simulating what happened at Cannae, but using the old breaking off rules is by far the easiest and least disruptive.

For, whether we call it a retreat, a break-off, a fighting withdrawal or whatever, the fact is that the Carthaginians DID do it. And yet the maneuver is IMPOSSIBLE to replicate in 2.2+, since Warband can't disengage from Blade except by a recoil result. Thus, there is no way for Hannibal to ORDER his Gallic warband to deliberately fall back and suck the Romans into his trap -- all he can do is hope that the Gauls lose melees without being destroyed in the process. Not a terribly impressive plan for one of the greatest commanders in history to come up with!!! This is why I believe it's fruitless to argue that such maneuvers are already simulated in the recoil rules.

The point I was trying to make is that a break off in DBA is not what you are describing. It leave the other player out of the ZOC allowing independent movement not directly ahead. I do agree that there is nothing in DBA that allows for a fighting withdrawal, absent an across the line recoil.

[Dangun]

Originally Posted by Kevin Boylan

although the consensus of professional historical scholarship on this topic definitely favors the ordered retreat theory

What consensus are you referring to?

Polybius does not contain any reference to Hannibal "ordering" a retreat. The closest you get is that Polybius writes that Hannibal planned to have the Romans trapped between the Libyans on either wing. This might imply that Hannibal deployed a soft centre, but says nothing about what he ordered during the battle. (3.115)

Both Polybius and Livy (22.47) suggest it was Roman action that drove back their enemy not anyone's orders.

DBA like most wargames probably grossly exaggerates the ability of any ancient general to change the course of a battle significantly once it started.

[Rich Gause]

Originally Posted by Dangun

What consensus are you referring to?

Polybius does not contain any reference to Hannibal "ordering" a retreat. The closest you get is that Polybius writes that Hannibal planned to have the Romans trapped between the Libyans on either wing. This might imply that Hannibal deployed a soft centre, but says nothing about what he ordered during the battle. (3.115)

Both Polybius and Livy (22.47) suggest it was Roman action that drove back their enemy not anyone's orders.

DBA like most wargames probably grossly exaggerates the ability of any ancient general to change the course of a battle significantly once it started.

The origional Tactica rules with all the heavy foot only being allowed to march stright forward probably had it closer to right. The Triarii actually being allowed to wheel made them elite! Not much fun to deploy and then no maneuver for a game though, ... I like DBA much better.