Wargaming the Battle of Zama

I’m interested in refighting the Battle of Zama. This is quite a popular focus for historical scenarios so I thought I’d have a look at what other people have done before collating my own thoughts. By chance, Zama was the theme of the Society of Ancients Battle Day in 2010. Not surprisingly the Battle Day, and Zama, got a lot of coverage in the Slingshot, the society journal, over the next few months. I’ve used those accounts to explore some key questions about how to refight the battle.

Key Questions

Various participants in the Zama Battle Day identified key questions that anyone organising a Zama refight needs to think about:

  • What was Hannibal’s game plan?
  • What troop types to use for the infantry?
  • How many Elephants models to use and how to space them out?
  • Whether to simulate, and if so, how to simulate the Scipio’s elephant lanes?
  • Whether to simulate, and if so, how to simulate the Roman’s manipular replacement ability?
  • How to ensure the third line of infantry (Veterans and Triarii) stay around to fight even if the rest of their army routs?
  • What to do about the 6,000 Numidian infantry on the Roman side?
  • How to deal with pursuing and returning cavalry?
  • How to make it a balanced game?
  • How to make it look like an ancient battle and achieve visual appeal?

Many of these questions were a challenge for the rules used in the refights. Leading to unhistorical results or, by those who thought ahead, the use of “Special Rules” on the day.

I’ll go through each of these questions, explain the problem, describe what various refights did to cope with that, and share my thinking about what to do.

What was the order of battle?

The Armati refight (Fry, 2010) gave each side the same number of units. Fry calculated the order of battle for the Roman army then gave the Carthaginians the same number.

Hayes (2010), for the Piquet (Archon 2) refight, picked the order of battle to give the armies a similar frontage.

For his Commands & Colors refight, Lawrence (2010) assumes the Carthaginians had 42,830 men and 80 elephants and the Romans 33,750 men. On a strict figure ratio this gave the Carthanginians more units but penalised them with fewer generals and less command cards.

Apparently the guys who put on the Civitates Bellatorum refight (MacDowall, 2010) had a debate on the size of the three Carthaginian infantry lines. MacDowall thought 12,000 in each but the other guy thought the numbers were weighted towards the third line because Scipio extended his infantry line before closing with them.

All Roman/Latin Legions/Alae in the Civitates Bellatorum refight (MacDowall, 2010) were four stands each of Velites, Hastati, Principes, and Triarii. That means a Triarii stand represented 150 men but the stands of the other lines represented at least 300 men.

Barnsdale (2010), who did the Peltast and Pila refight, agreed with MacDowall (2010). He believes Polybius was exaggerating Carthaginian numbers. The consequence for the game was a reduced Carthaginian 3rd line of Veterans.

Hey (2011) believes the Roman army had both more infantry and more cavalry than the Carthaginians. He argues this because:

  • Unusually, Polybius is relatively silent on the respective orders of battle, which Hey interprets as being shy of belittling Scipio’s achievement
  • Hannibal’s early battles were against numerically superior Roman forces
  • The Romans had control of the sea so could ship pretty much anything they wanted
  • Hey (2011) assumes Hannibal brought less than 10,000 men back from Italy including no cavalry
I’ll mull this over a bit more. Generally I find the arguments put forward by Hey (2011) quite compelling.

What was Hannibal’s game plan?

Mark Fry (2010) thinks Hannibal was trying to fight an “attrition battle”. Wear the Romans down with his first two lines and use the third line for the winning punch.

“Attrition battle” or envelopment that is the question. I suspect he was going for both.

What troop types to use for the infantry?

What troop type to use for the Carthaginian foot, specifically the Gauls, Ligurians, Carthaginians, Libyans and “Veterans”

For his Armati refight, Fry (2010) believed the Roman propaganda and used the used standard classifications from the rules to implement it:

  • Roman:
    • standard Armati classifications but upgraded the Triarii and downgraded the Allies.
    • They were deployed in three continuous lines, without gaps, but the third line was half the width of the other two (Heys, 2010).
  • Carthaginian:
    • First line: Gauls (Warband), Ligurians (Javelin Light Infantry), Balearic (Sling Light Infantry).
    • Second Line: Libyan and Carthaginian (Levy).
    • Third line: half African (Veteran) and other half Italians and Spanish

The Commands & Colors refight (Lawrence, 2010) didn’t buy into the Roman propaganda nor the establish deployment for the Carthaginians:

  • Roman: Hastati (Aux), Pricipes (MI), Triarii (HI)
  • Carthaginian:
    • First line: Gauls (Aux rather than Warband), Ligurians (not mentioned so presumably bundled as part of the Light Infantry or perhaps the Gauls), Balearic (Light Infantry).
    • Second Line: Italians (MI) and Spanish (Aux).
    • Third line: Veterans (HI) and local levy (MI)

The Civitates Bellatorum refight (MacDowall, 2010) had the Romans as “A Grade” and the Carthaginians heavy infantry ranging from “C Grade” up to “A Grade”:

  • Roman: Hastati and Principes (A Grade armoured, pilum swordsmen), Triarii (A Grade armoured, understrength spearmen)
  • Carthaginian:
    • First line: Gauls and Ligurians (C Grade javelinmen), Balearic (C Grade skirmishers; he also adds in Libyan light infantry and Moors).
    • Second Line: Punic Infantry (B Grade javelinmen).
    • Third line / Reserve: half African Veterans (A Grade armoured spearmen) and other half Bruttian Veterans (A grade armoured swordsmen)

Brian (2010) included an orbat for Impetus. But I don’t understand the Impetus codes so won’t reproduce it here. I understand from Heys (2010) that the Roman infantry consisted of three lines of four groups of heavy infantry separated by small gaps in each line and an interval between th elines.

Barnsdale (2010) mentions that in his Peltast and Pila refight he:

  • made the “battle hardened” Gauls and Ligurians “Regular” rather than “Levy”
  • reduced the size of the Carthaginian veteran line

The DBMM refight (Partridge & Mather, 2010) was another one that believed the Roman propaganda:

  • Roman: Hastati and Principes Reg Bd(O) and Triarii Reg Sp(S)
  • Carthaginian:
    • First Line: Celts Irr Wb(O), Ligurians Irr Ax(O), sundry Psiloi
    • Second Line: Citizen and Libyan Infantry Reg Sp(I)
    • Third Line: Veterans Reg Sp(S)

The Piquet (Archon 2) refight (Hayes, 2010):

  • Roman: 18 units of foot represented five legions and had a frontage of 4 feet
  • Carthaginian:
    • First Line: Competent
    • Second Line: Competent
    • Third Line: Competent and experienced. Two units of elite.

In the AWE refight (Silva, Hills, Jerboa Wargames, 2010):

  • Roman: Hastati and Principes EdH and Triarii ??
  • Carthaginian:
    • First Line: Presumably LyM (close order infantry with mixed weapons) and BwL (archers and slingers) and SpL (javelinemen)
    • Second Line: LyM (close order infantry with mixed weapons) and SpM (average close order Spearmen)
    • Third Line: Veterans Reg Sp(S)
Personally I reject some of the Roman propaganda so for the Carthaginians:

  • First Line: Both Gauls and Ligurians should be competent tribal warriors
  • Second Line: Despite the bad press about these guys they gave the Romans a hard fight so I think they should be competent rather than levy.
  • Third Line: Veterans are veterans. As good as the best of the Romans.

How many Elephants models to use and how to space them out?

Hannibal probably spread his elephants across his army’s entire frontage. That is quite hard to simulate in a table top game because it implies a very low density of elephants per metre in the front line. Refights tend to assume skirmish infantry were in the elephant line.

The Armati refight (Fry, 2010) assumed Hannibal mixed his elephants with his light troops. You can’t do that in Armati so Fry classified the elephants as light troops for deployment but heavy troops for combat. Fry also made them “Non-key troops” so they wouldn’t count for morale when they routed. The Carthaginian order of battle included eight Elephant units supported by nine skirmisher infantry units – the total (17) is a lot more than 12 units in the Carthaginian 3rd line.

The Commands & Colors refight (Lawrence, 2010) had two elephant units in front of the main battle line along with one light infantry unit. Quite thin compared to the six units in the Carthaginian third line.

The Civitates Bellatorum refight (MacDowall, 2010) had the elephants spread along the front, with some in each of the four commands in the front line. Previous play test games showed, without the elephant support, the Carthaginian horse were walked over by the the Roman cavalry. There were 12 elephant stands in total supported by 22 stands of skirmishers. So 34 stands in the elephant line compared the to the 40 stands in the 3rd infantry line (the largest of the lines).

In the Impetus refight (Brian, 2010) there were four units of elephants to seven units of skirmishers. Contrast with five units of veterans (one of “Hannibal’s Veterans” and four of “Vet Libyan Infantry”).

The DBMM refight (Partridge & Mather, 2010) had seven elephant stands alongside six Psiloi. Which isn’t much compared to the 34 elements of Veterans in the third line.

The Piquet (Archon 2) refight (Hayes, 2010) gave the Carthaginians 12 models in three units. Apparently this number was selected to “match” the Roman frontage. However judging from the map there were gaps between the elephant units.

Hey (2010) observes that “although Zama is strongly associated with elephants, it seemed that their involvement in the refights was a lengthy distraction in most games, out of proportion to their importance.” He cites the DBMM game as an exception; the Roman velites had destroyed all the elephants by the time Hey saw the table.

It would be nice if the combined elephant and skirmisher line extended the entire frontage of the Carthaginians.

Whether to simulate, and if so, how to simulate the Scipio’s elephant lanes?

Heys (2010) expected to visually see the elephant lanes on table, and in his review of the various games noted their absence as a deviation from history.

The Civitates Bellatorum refight (MacDowall, 2010) had a “Special Rule” for the elephant lanes: “in an inconclusive combat the elephants would pass through the Roman formation”.

Hayes (2010) introduced a elephant lane “Special Rule” for the Piquet (Archon 2) refight. The Roman player could adopt the chequer board shape of the triplex acies or the “lanes” formation. In both cases units had gaps between them. In the “lanes” option the elephants had to pursue down the lanes until they popped out the back, without fighting the lines they were passing.

Hayes (2010) did consider an alternative “Special Rule” where the Romans get a “plus 1 for being trained against elephants” but rejected that as too bland.

I disagree with Heys (2010). I believe the manipular formations, whether triplex acies or elephant lanes, are inside the on-table units of a particular Roman line, so do not need to be visually represented.

However, the elephant lanes were a key part of Scipio’s strategy for this battle so should be represented somehow. I quite like the “Special Rule” of the Civitates Bellatorum refight (MacDowall, 2010) that meant “in an inconclusive combat the elephants would pass through the Roman formation”.

Whether to simulate, and if so, how to simulate the Roman’s manipular replacement ability?

Apparently the Quincunx formation is built into the Civitates Bellatorum rules (MacDowall, 2010). One day I’ll download them and see how they do it. In the photos of the game each Roman element seems to be separated from all of its neighbours.

“Line Relief” for Roman Legions is also built into Impetus (Brian, 2010) and this feature was used/attempted extensively in the refight (Heys, 2010). Apparently “Triarii” and Hannibal’s “Veterans” also have special abilities under Impetus.

The Armati refight (Fry, 2010) had a “Special rule” to simulate manipular replacement:

  • Allows Roman Legionnaire units to interpenetrate each other without penalty.
  • Makes Roman Legionnaire units with another such unit behind resilient in the face of units with impetus (Elephants, Warband or Heavy Cavalry)
I believe it was the resilience of the Roman infantry that won their battles, not any benefits from weapons and armour. The triplex acies formation, if it existed at all and their those that doubt it, would have contributed to this resilience because it is assumed to be part of the manipular replacement system.

However, line replacement was explicitly absent from Zama. I believe the use of the elephant lanes prevented the Roman infantry from benefiting from their normal line relief.

How to ensure the third line of infantry (Veterans and Triarii) stay around to fight even if the rest of their army routs?

Hampshire (2010) used standard WAB rules for his refight. That meant the the Carthaginian second line routed when they saw the first line rout. Not very historical.

The Armati refight (Fry, 2010) radically changed the morale rules:

  • Third line exempt from routed into tests.
  • Only troops in third line were “key troops”.

The Impetus refight (Brian, 2010) increased the morale contribution of the rear lines (from VD 3 to VD 4 per unit) to ensure the heavy weight troopers got engaged before the 50% VD demoralisation point was reached.

Sabin (2010) points out that the Lost Battles rules have two mechanisms built in that ensure the third lines will remain to fight. Firstly, armies become “spent” without anyone breaking. Secondly, panic is calculated on a unit-by-unit basis, rather than a ‘threshold’ which most of the other rules seemed to do.

The DBMM refight (Partridge & Mather, 2010) used DBMM’s “variable Morale Equivalence” to ensure the veterans hung around. The elements of the third line had 2ME and those of the first line 0.5ME. The elephants were classed as expendable so had 0ME. Hannibal’s command, including a huge complement of veterans and the cavalry, had over half the ME of the army in total.

In the AWE refight (Silva, Hills, Jerboa Wargames, 2010) “the lines would fight and break independently”.

Something that is very specific to the rule set used. Although several refights fiddled the army break point to achieve the effect I think the AWE refight (Silva, Hills, Jerboa Wargames, 2010) had it right: the lines should fight and break independently.

What to do about the 6,000 Numidian infantry on the Roman side?

Livy says Masinissa joined Scipio before the battle “with a force of 6000 infantry and 4000 cavalry”. Then the sources go silent on the Numidian infantry. They either refer to the “Numidians” as a whole or to the “Numidian horse”. For example Livy says that on the day of battle “Masinissa and his Numidians being posted on the right”. Appian says Scipio had “had as allies Massinissa with a large number of Numidian horse, and another prince, named Dacamas, with 1,600 horse” and that Scipio “stationed his Numidian horse on his wings”. Polybius says of the Roman deployment “on the right wing Massanissa with the whole of his Numidians”

Where were the Numidian infantry? Did they play a part in the battle at all? If they do, how do you ensure they don’t unbalance the game as this is a significant proportion of troops available to the Romans?

Apparently most of the refights at the SOA Battle Day did not represent the Numidian infantry (Hey, 2010). For example, there were no Numidian infantry in the Impetus refight (Brian, 2010).

However some did include the the Numidian foot.

The Armati refight (Fry, 2010) classified the Numidian infantry as javelin skirmish infantry, so could support the Numidian cavalry when in melee.

The Commands & Colors refight (Lawrence, 2010) only included 4800 Numidian infantry and classified them as LI.

In the Civitates Bellatorum refight (MacDowall, 2010) the Numidian infantry were outnumbered by the cavalry. 32 stands of Numidian cavalry versus 16 of infantry. They is equivalent to only 2,000 men assuming each stand is 125 men based on the cavalry numbers. The infantry were classified as “C Grade skirmishers”. And interestingly MacDowall spits the Numidian infantry between the two wings so half of them were stationed with the Italian horse.

Sabin (2010) includes all 6,000 Numidian infantry in his Lost Battles refight. Their classification as “Levy” means that they contributed only 2% of the Roman fighting value but 17% of the manpower.

The DBMM refight (Partridge & Mather, 2010) included 12 elements of Numidian infantry under Massinissa, eight Ps(S) and four Ax(O). The latter are Roman trained infantry; Duncan Head (2010) also says that survivor’s of Syphax’s trained infantry could have been present. Partridge and Mather opted for a non-standard ratio of men to elements (400-500) so this corresponds to the 6,000 men you’d expect. Which is where it gets a bit odd because Massinissa has nine elements of cavalry LH(O); at the nominal scale that would be 4,500 men and he only had 2,000.

I think the Numidian infantry should be present; just because the sources don’t mention them doesn’t mean they didn’t fight. I also find it intriguing that having Numidian “heavy infantry” is a possibility.

How to deal with pursuing and returning cavalry?

Reading the battle report of the Civitates Bellatorum refight (MacDowall, 2010) it is clear that the rules cover cavalry pursuing off table and then returning.

The Armati refight (Fry, 2010) introduced a “Special Rule” for “victorious” Cavalry forcing them to lurk on the rear table edge for a while.

Cavalry rushing off the battlefield and then, sometimes, returning latter is a key part of battles in this period and must be simulated. Either as part of the core rules or as an add-on “Special Rule”

How to make it a balanced game

I covered the Order of battle above and different refights made completely different assumptions. For example the Commands & Colors refight (Lawrence, 2010) gave the Carthanginians more units but penalised them with fewer generals and less command cards. Whereas the Armati refight (Fry, 2010) assumed equal numbers of units.

What the above doesn’t address is how to prevent the typical Roman mincing machine of most rule sets? Hey’s (2010) observation was that in most of the refights the Romans “usually managed to chop their way through regardless. This was especially the case when the hastati met, and made mincemeat of, the Carthaginian Citizen infantry which was apparently an equal content.” Taylor (2010) says the Romans won 10 of the refights and the Carthaginians won only four.

In the Peltast and Pila refight (Barnsdale, 2010) the “legionaries were cutting through the Carthaginian citizens and the Ligurians like a knife through butter”. Apparently Tony Bath was pro-Roman and his rules reflect that.

In the WAB refight (Hampshire, 2010) the “Roman infantry seemed even more crushingly superior than one would expect”. The Carthaginian first line “bounced off the Hastati and fled on contact”. “In fleeing they also caused a fair bit of panic; so they second line didn’t really get to fight as a line.” Finally, contact with the third line “saw the Romans sweeping all before them”.

Lost Battles (Sabin, 2010) doesn’t seem to suffer this problem. The Lost Battles team refought Zama three times during the SOA Battle Day. The results were two Carthaginian victories to one Roman victory.

I think most rules over-rate the Roman heavy infantry compared to their contemporaries. I don’t see the meat grinder effect of most rule sets in the historical accounts. Personally I think it was their refusal to accept defeat, both tactically and strategically, that made them so successful. They had a resilience that other nations lacked, and it is this resilience that needs to be simulated, not pila, armour or the triplex acies formation.

How to make it look like an ancient battle and achieve visual appeal?

The Armati refight (Fry, 2010) used “Intro scale” (a unit is one element) rather than the more normal Optimal scale (a unit is four elements). However, Fry made all heavy infantry units two bases deep to make them look better.

Hey (2010) observed of the WAB refight that “both armies were arranged in lines of blocks of figures with substantial gaps between the blocks (!). Thus there was no impression of any solid lines in Hannibal’s army and no sense of any lines of battle.” Shame.

Apparently the folks behind the AWE refight (Silva, Hills, Jerboa Wargames, 2010) do not believe in entire flat tables. So after consulting Google Maps for “Saqiyat Sidi Yusuf, Tunisia” they placed an area of “rough” on the north-west portion of the table.

The Impetus refight didn’t go for visual appeal (Brian, 2010). The visual spectacle, such as it was, was provided by a plain tan base cloth.

Simon Miller's Zama
Simon Miller’s Zama – 28mm Command & Colors

In contrast Simon Miller (2010) pays particular attention to the aesthetics of the game. Aside from the beautiful 28mm figures painted by an army of volunteers, Simon also lovingly crafted a custom made Zama table. 16 feet by 4 feet of North African glory. On the terrain boards Simon modelled fields (both ploughed and with standing crops), tracks, ditches, hay stacks and trees. I suggest you check out the photo’s of Simon’s impressive Operation Zama.

I’ve had the pleasure of fighting on Simon Miller’s table. And it looks fantastic. But I’m more a modular kind of guy so a generally tan base cloth with a bit of rough in the north-west will probably do me.


Brian, M. (2010, July). Must Triarii Harder – how Roman wasn’t Destroyed in a day. Battle Day using Impetus. Slingshot, 271, p. 28-29.

Barnsdale, D. (2010, July). Zama with Peltast and Pila. Slingshot, 276, p. 33-34.

Fry, M. (2010, July). Zama with Armati. Slingshot, 271, p. 17-22.

Hampshire, S. (2010, September). WAB at Zamaday 2010. Slingshot, 272, p. 22.

Hays, A. (2010, September). Organising the Piquet (Archon 2) Game at the SoA Battle Day. Slingshot, 272, p. 20-22.

Head, D. (2010, September). Zama Battle Day Questions. Slingshot – The Journal of the Society of Ancients, 272, p. 12.

Hey, S. (2010, November). Battle Daze 2010: Zama and how it went on the Society’s Battle Day. Slingshot, 273, p. 31-32.

Hey, S. (2011, May). What Really Happened at Zama? Slingshot, 276, p. 9-20.

Lawrence, A. (2010, July). Zama with Commands & Colors: Ancients Epics. . Slingshot, 271, p. 22-25.

Lockwood, R. (2010, January). Zama 202 BC – The 2010 SoA Battle Day Battlepack. Slingshot, 268, p. 2-5.

MacDowall, S. (2010, July). Civitates does Zama with Very Small Men. Slingshot, 271, p. 26-27.

Miller, S. (various). Operation Zama. Big Red Bat.

Miller, S. (2010, July). Operation Zama. A very large 28mm Command & Colors: Ancients game. Slingshot, 271, p. 31-33.

Partridge, T. and Mather, D. (2010, September). Zama in DBMM. Slingshot, 272, p. 15-19.

Sabin, P. (2010, July). Zama with Lost Battles. Slingshot, 271, p. 30-31.

Silva, A., Hills, J. and Jerboa Wargames. (2010, September). Zama with AWE – Arcane Warfare Excel. Slingshot, 272, p. 23-24.

Simpson, M. (2010, July). Zama with Donald Featherstone’s Rules for Ancient Wargames 300BC to 400AD. Slingshot, 276, p. 35.

Taylor, R. (2010, September). Zama Battle Day – some thoughts in conclusion from the Editor. Slingshot, 272, p. 24.

Leave a Reply