Campaign battle screenshots and videos

[Geoffrey of Villehardouin]

Second Siege of Gul'dursun - short SS6.1 Khwarezmian campaign, 1222 AD

The Mongols east of the Amu Darya had been defeated and had lost their two generals and most of their infantry. Their morale had plummeted but they were still strong and seemed determined to cross the river to join their Khan. So a follow up to their first attack on Gul’dursun was to be expected.

The two forward river forts were reinforced with fresh troops and when Gul’dursun was predictably attacked, its garrison had not one but two companies of Hashashim and not three but four full strength companies of spearmen - and of a higher quality than those of the previous battle. The Hashashim companies had been retrained at the Hashashim HQ at Urgench and had two chevrons of experience. They were going to be crucial in lowering the morale of the leaderless Mongols. On the walls they would be invincible.

So, compared to the previous battle, this was going to be an easier task, though the odds on paper favoured again the Mongols. The strategy still was to delay the ram with all kinds of ranged fire but only long enough for the siege tower or ladders to be placed on the walls, so that the Hashashim could take care of the Mongol foot archers on the battlements and the spearmen could deal with the Mongol cavalry pouring through the gate.

Everything was timed to perfection and the battle went according to plan. The defenders fought with the best possible tactical advantages over their foe and the knowledge that a victory this day would end the Mongol threat.

May the memory of the Khwarezmian warriors still live in marble when the bones of the Mongol dead have turned to dust.

-----------------------------------------------------------------------------------------------------------------------
[FKoC AAR] Times full of Distemper
Reviewed by robinzx at the Critic's Quill, Issue 31
[BC AAR] The Maharajah and the Guild of Thieves – a Chauhan Rajput AAR
[1648] Thirty Years' War
[Kingdoms AAR] Antioch Crusaders Mod campaign

[Geoffrey of Villehardouin]

Stainless Steel pike & shot formations

Ingenious as Total War is, it is sometimes easy to forget it is a game and that the AI is not even human and therefore has its faults. This is especially apparent in the late period battles. Stainless Steel is probably the only mainstream mod that offers the possibility for pike & shot warfare in the late period. There are units of even the later 16th century. Unfortunately, the same defects one sees in the battle AI behaviour in For King and Country and in 1648 Dreissig Jahre Krieg are present also in Stainless Steel. Even more so in Stainless Steel, because in the two total conversion mods for 17th century warfare, the medieval battle AI has been slightly adjusted to behave more appropriately.

In Stainless Steel, the AI formation starts well but organisation disintegrates as soon as the AI units come to within musket range. The shot companies get in the way of other shot companies, then halberdiers get in the way of everyone and finally pikemen repeat the act, so that with larger battles it is impossible for the AI shot companies to have a clear line of fire. All the while the computer player's units are being pounded by the player's artillery and shot companies. Moreover, the AI artillery is placed at the back of the computer player's army, as if it was medieval artillery, and has little chance to contribute.

Below are two videoclips of two late period battles. You can watch them in HD, at 720p or 1080p depending on your internet connection, by clicking on the gear icon at the bottom of the YouTube videoclip and adjusting the settings accordingly.

In this first example, the Spanish army has set up in two Spanish tertios. The offset arquebusiers are protected by cavalry in the centre and there is also some cavalry protecting the flanks. There is also a ribault unit in each army. Both sides have practically identical armies other than that the Milanese (Genoa) are a little heavier overall, so the odds should favour them.

Unfortunately, the player can only too easily defeat the AI. The aggressiveness that served the AI well in the early medieval periods is its downfall in the late period. That was probably true in real history, that was why the later armies are sometimes called "professional". It is unfortunate the game does not seem to have the tools, the code, to correct the aggressive behaviour of the AI.

In the example in the videoclip that follows, the Papal States are attacking a Venetian army set up in pike and shot formation with musketeers at the front and some arquebusiers in the rear. The Papal States army is stronger on paper but the Venetians have several monster ribaults. The Papal States cavalry attacked the flanks and were repulsed. Then the infantry simply walked and walked and walked, within the firing range of the monster ribaults and the musketeers, until they had been cut to pieces. It was a wonder anyone was killed on the Venetian side.

In both these battles, I gave practically no commands until most of the AI army (other than the artillery) had been routed.

In previous posts you can find examples from the 1648 mod with several pike and shot formations, where at least the AI seems to have something of a chance. The fault of the AI in all these cases was that they had for most of the battle the shot behind the hamberdiers and pikemen and the artillery behind the entire army - rather than at the very front. The melee units, advanced too far ahead, in front of their musketeers and arquebusiers. In Stainless Steel, moreover, the shot units did not stop to walk until they had come too close to the human player's army, by which time the shot companies were also hopelessly mixed up with each other and unable to get a clear line of fire. Bunched together like that, they were massacred by the Venetian monster ribaults and musketeers.

What is ideally needed is a completely different battle AI, specific for the later period, that will space out the crossbow and shot units and prevent the AI melee units from walking in front of them. The role of pikemen was defensive in that period, inspired by late Roman military manuals that divided the troops into cursores (missile troops) and defensores (the spearmen). The spearmen were there only to fend off enemy attacks, e.g. by the enemy cavalry. They were generally not supposed to attack themselves. In the period of pike and shot formations, only cavalry attacked and that only on the flanks. Cavalry elsewhere, as in the German pike and shot formations, was intended to provide additional protection to field artillery and shot companies between the pike blocks. Its role was also defensive. The current battle AI is ok for early undisciplined medieval armies, where archers can still shoot overhead, but it is unsuitable for late period armies.

[Geoffrey of Villehardouin]

Funny things the AI does

Despite the continual efforts of modders and the variety of battle AIs used in the various mods, and even though the currently available AIs are much improved compared to the original, the battle AI still has a knack for getting things wrong. Here are some funny things they AI does. They are funny in a painful sort of way, like when you are watching someone making bad things worse.

This videoclip is from retakes of the previous custom battle (most scenes are from a single retake), while I focused in on the mistakes the AI did. The first mistake was that it moved unnecessarily close to my pike and shot formation. They could have started firing from much much further away. Then there were nine other incredible things the AI did. Watch them here:

You can watch the videoclip in high definition by setting the resolution to 720p or 1080p depending on the speed of your connection, by clicking on the gear icon at the bottom of the YouTube videoclip window.

I have tinkered with the AI myself and impressive as it is in making battles happen, it nevertheless seems impossible to stop from taking some really awful decisions. They keyword here is "stop". You really cannot stop things easily, when it comes to the battle AI. Most script seems to be about its priorities in what to attack, attack, attack.

[Geoffrey of Villehardouin]

Battle tactics tests

One instinctively expects to see the battle AI fighting in some cohesive way, like a real army would. Most of the time, however, any semblance to a disciplined army disappears as soon as the computer player advances to join battle. By frustrating that advance by placing various advance parties or a few units close to its flanks, the AI army is delayed and the more it is delayed the more it has a chance to fall apart into a random mass or for its units to run off in random directions. In these two videoclips, two small groups of infantry were sent forward one on each flank of the advancing AI army. Watch the AI army disintegrate.

I tested vanilla and several mainstream mods covering the medieval period. As usual, in these battles both armies had close to 20 units. You can see at the start the battle order for the human army. The computer player had a very similar army, well balanced, but with slightly heavier units. So in both cases the AI army was at least as strong in all sectors. Other than for the general's bodyguard, each army had only one heavy cavalry unit. I wanted to test how well could the AI fight what would be an infantry-heavy battle.

The first example is in vanilla. In fact in this battle I gave relatively few commands other than for the three units on the right wing "forlorn hope" - two infantry and one cavalry unit. The AI basically lost fighting a stationary army.

At first it sent its strongest cavalry into a suicide charge on the right "forlorn hope". Once pinned down, they were charged in the rear by my own single unit of heavy cavalry and were routed. Subsequently, that single unit of heavy cavalry had a free rein to charge a couple of AI infantry on that side. At that point (minute 1:35), with only just a couple of units routed, the AI army, that was already scattered, milling around with little sense of direction, simply panicked and ran.

The second example is from a Stainless Steel custom battle, fantasy late Byzantines vs fantasy late Sicilians/Normans. At the start you can see the deployment, and what I mean by forlorn hopes - the two forward placed groups on the flanks. With the SS AI, battles are tougher, and so I gave more commands to the units at the two flanks. The centre remained stationary. The main cavalry unit, with my apologies, was a mercenary Sicilian cavalry unit and looks like the Sicilian units the AI commands, other than that it has a Byzantine flag. Placing the two forlorn hopes at the flanks caused a complete and utter disintegration of the Sicilian/Norman lineup. Units wandered off in random directions. Subsequently, that one mercenary cavalry unit routed the entire AI army.

You can see the loss in cohesion already at 0:55 and by 2:15 many AI units seem bunched up on the upper right-hand corner but not in any kind of proper formation. By minute 2:30 those that have not been routed are milling around in random directions.

Cases like these make me wonder if, had the AI army done nothing - if they had not moved or marched anywhere, just sat somewhere in a defensive stance with their flanks protected and did not move at all - you might then get a more balanced battle.

The battle AI has no apparent code for how to keep the army together and no code for assessing the overall situation on the battlefield. As someone in the Shogun 2 section commented, the paper-rock-scissors model for who to attack causes the AI to fight as in a game of chess, with units moving in every direction. wherever you put for them a target for attack, rather than as in real armies. I find this most apparent in vanilla and Stainless Steel. Of course battles become a little unrealistic when you put spearmen in the forlorn hopes, the AI sends swordsmen, who you can then run down with cavalry kept on reserve. Or you can put also some swordsmen in the flanks, as in the demonstrations above, and send the enemy spearmen into a mary-go-round. Or if their army is mainly swordsmen, you can send some cavalry in the flanks and they will disintegrate. The worst of it is that the AI having zero, absolute zero code available to help it use pike and shot formations in the late periods and indeed it is a swine to even get the AI to use firearm units effectively. They get in each other's way and behind the pikemen, as if they were archers and could shoot overhead. The code seems to be inherited from Rome and AI armies fight rather too much like complete barbarians. The unit spacing code on deployment does not work once the battle has started and deployment is anyway useless, if the AI units cannot keept it during the battle.

Of the various mods I have tried, Deus lo Vult and Rusichi Total War performed best in infantry battles such as the ones demonstraed above. The tactic of the forlorn hopes on the two wings did not work at all in Rusichi Total War. They were attacked, but the AI kept its centre in relative cohesion. In fact, I could not get the Rusichi TW infantry to scatter as much, so this disintegration of the AI army is not completely inevitable.

[Geoffrey of Villehardouin]

Greek fire again

Greek fire works best in sieges. However, unlike in real history, Greek fire units cannot fire well from walls at those below. They can only fire at units at the same level as they are, which is maybe not completely historical and makes their use very risky.

These are three tests of using Greek fire in gate defense. In the first two tests, they are protected by pikemen who help keep the enemy horsemen at a distance. It usually worked. In the third case, they are only supported by militia SS6.4 Contaratoi. The odds were 1:5 against them. But it was still possible to win, as long as the enemy cavalry did not charge straight at them.

In all these tests, the defenders had two units of flamethrowers, however, in the first two tests the second unit, which was kept on reserve, did not see any action.

You can watch them in high definition by clicking on the gear icon and selecting 720p or 1080p, depending on your internet connection speed.

This tactic would not probably work in most other mods, where flamethrowers are firing more slowly.

[Geoffrey of Villehardouin]

Two short throwaway battle videos with some quasi-atmospheric scenes from battles in the Po valley in the early part of a late Venice campaign (Stainless Steel 6.4). The Gracul AI throws no end of armies at you in interesting locations and makes sure you have to attack them.

A battle near Mantova against some very unusual rebels



Fourth Genoese attack on the fort of Pavia

[Geoffrey of Villehardouin]

The graphics are probably part of what make Total War battles addictive. Here are some atmospheric views of Stainless Steel Adana, with apologies for the slow frame rate.

The Seljuk Turks saw a relatively small garrison and decided to declare war over Adana. Unfortunately for them there was a strong army in Cyprus that was ferried over in good time. The computer player must have regretted starting a war so inauspiciously.

My apologies again for the slow frame rate. Hopefully it does not ruin too much the atmosphere. Imagine watching events back in time through a revolving gyroscope. Below are some screenshots from this battle.

Spoiler Alert, click show to read: 

[Geoffrey of Villehardouin]

Chivalry II - the Sicilian Vespers Byzantine campaign, early period

The Chivalry II mod has in my view the most historically accurate Kontaratoi (with relatively long spears), Scoutatoi and Menavlatoi (armed with a short pike-like menavlon). They are available fairly early on, as they should, since they had been standard units since the 6th century or so. Their skins are also quite nice in my opinion and with the additional units from the CBUR project make the Byzantines a very attractive faction. Unfortunately, Chivalry II is meant only for the high period, so the starting positions in the early period are a little strange and involve the Normans already in Greece in 1098 but more strangely also the Moravians (?) and Franks. Also strangely, the Mongols are already in Caucasus and Anatolia. But the temptation of playing the Byzantines with what seemed potentially a historically accurate roster was irresistible.

For this campaign, their leaders have been modded so as to have not only Ioannes Ducas and Nikephoros Bryennios but more importantly Georgios Palaeologus among their generals and Anna Comnena as a princess. Georgios Palaeologus was probably the first well known member of that family which was later to produce so many Byzantine emperors and was the best known general of Alexius Comnenus, so indeed it was a bit of a surprise he was not in the initial roster and I had to mod him in.

The starting position of the Byzantines was not what one would expect for this period, as they only controlled Constantinople, Adrianople, Thessalonica and Mistra. Not only that but the rest of the Byzantine lands were controlled by no less than 8 other factions - Seljuks, Genoa, Venice, Moravians, Latin Empire (Franks), Normans, Serbs and Bulgarians, almost in all cases unhistorically. Making a start at recovering the starting position of the historical Byzantines in the 11th Century would have involved going to war against 8 factions. I took the more conservative option of allying with the Serbs, allowing the Bulgarians, Venetians and Genoese to keep their settlements and going to war with the other factions but only one at a time, starting naturally with the Normans who controlled Athens. Having started something, one begins to think about finishing it, which led to a landing of Georgios Palaeologus in Southern Italy to recover the Byzantine lands there and bring the fight to the Normans, as the Comneni attempted to do on two occasions. Palaeologus captured Cusenza which provoked several counterattacks and sieges by the Normans. One of these is shown in the next video.

Historically, Palaeologus had been charged with the defense of Durrhachium against the Normans, which he accomplished successfully for about a year. In that time he placed catapults and ballistae on the ramparts to defend against the Normans, built a siege tower inside the city to oppose the siege tower the Normans built outside, equipped the men on the ramparts with naphtha incendiary grenades to hold back the attackers and made several sallies in one of which he fought apparently with an arrow embedded in his head, finally succeeding in setting the Norman siege tower alight.

In this siege, his men sallied out with catapults, since they could not be placed on the ramparts, and bombarded the Normans who mysteriously, having first laid siege on Cusenza, hesitated now to attack. Indeed they only attacked after they had been harassed by the Byzantine mounted archers. Their attack turned out a complete debacle, attacking as they did piecemeal with only some cavalry and a unit of dismounted knights ever making it to the Byzantine lines, were they were every time quickly routed. The Byzantine position was unassailable with spearmen supported by swordsemen and several units of archers and javelinners offering support from the bastions. The light cavalry picked up the routers.



Screenshots from the siege of Cusenza

Spoiler Alert, click show to read: 

But the Normans did not give up and indeed succeeded once in retaking a fort in Southern Italy, in the approximate location of Byzantine Skyllicion (Squillace), that the Byzantines had captured after a successful siege.

Screenshots from the Norman siege of Skyllicion

Spoiler Alert, click show to read: 

They also laid again siege on Cusenza, which was lifted by the appearance of another Byzantine general, Andronicos Agiopantites, who landed with reinforcements from Greece. Indeed for the entire campaign on Italy the Byzantines had to depend on line infantry to be ferried over from the east because it seemed impossible to raise decent units locally, save from rare occasions of some rather expensive crossbowmen, heavy spearmen and mounted milites becoming available in Southern Italy or Sicily mostly as mercenaries.

In any case the combination of Menavlatoi, Contaratoi and Scoutatoi ferried in from Arta combined with ranged units from Cusenza, that included some modded naphtha grenade throwers, allowed Andronicus Agiopantites to use time honoured cursores-defensores Byzantine tactics to defeat the Normans in an open battle. The Normans, as expected being controlled as they were by the AI, fought with little coordination but surprisingly at first avoided attacking and when eventually were persuaded to attack, succeeded at one point to bring most infantry into the battle, while their cavalry sensibly avoided to attack the Menavlatoi. Nonetheless, the AI as usual failed to maintain its initial deployment, or to use guard mode or allow a field of fire for its crossbowmen, and seemed to pay little regard to the fact that it was usually locally outnumbered due to the depth of the Byzantine deployment and the support of Byzantine spearmen by ranged units. The result was a crashing victory for Agiopantites.



Screenshots from the battle

Spoiler Alert, click show to read: 

The Normans withdrew to a defensive stance around Messina, and never regained the initiative in Italy.

-----------------------------------------------------------------------------------------------------------------------
[FKoC AAR] Times full of Distemper
Reviewed by robinzx at the Critic's Quill, Issue 31
[BC AAR] The Maharajah and the Guild of Thieves – a Chauhan Rajput AAR
[1648] Thirty Years' War
[Kingdoms AAR] Antioch Crusaders Mod campaign

[Geoffrey of Villehardouin]

Roundaway Hill

This is another battle from my AAR, Times full of Distemper. It is the equivalent of the Battle at Roundway, fought at approximately that time in a nearby location by approximately the same generals.

In the campaign battle, Alexander Hopton's Foote marched out of Devizes, where they had barricaded themselves, and joined up with the Horse that had arrived from Oxford. The Royalist Foote attacked first, sending the musketeers forward in the usual ill-advised AI tactic. The centre was under the command of William Waller. Under AI control was the Parliament Horse on the left and Robartes with the reserves in the rear.

William Waller had little besides his own bodyguard for cavalry, so when the Royalist musketeers skirmished forward, he made a few charges with his bodyguard and spread havoc in those companies. Because of the losses at Lansdown, there were only two and a half companies of Parliamentarian pikemen. Somehow the AI decided to concentrate the attack against the location where Waller stood, near some of the pikemen. There was so much Royalist Horse the Parliamentarian pikemen were overran and with them the musketeers. Robartes and Haselrig, being under AI control, took forever to enter the battle. How was this battle won, it was a miracle. Had Haslerig and Robartes delayed half a minute longer it would have been a dead loss.

In the historical battle, things went even worse, even though the Cornish Foot stayed in Devizes until the very end and took no part in the more crucial stage of the battle. The Royalists therefore, had only Horse and two light cannon (drakes, lighter than sakers). The battle was fought in very open fields, which clearly favoured the Royalist Horse. It was a disastrous defeat for Parliament. How could William Waller have made such a bad mistake after his successful defense at Lansdown?

To begin with, the Cornish Foot, all the army that the Royalists had West of Oxford, had been barricaded in Devizes, with their general Ralph Hopton badly wounded. If they surrendered, William Waller would have achieved a stupendous victory that could have decided the war. So understandably, he was willing to take his chances. The Cornish Foot were greater in number (about 3000-3500 compared to Waller's maybe 1500 ínfantry plus about 1000 or so Horse) but they were demoralised after the loss of Bevil Grenville and with Ralph Hopton wounded. They had begun negotiations for their surrender, whether with sincere intentions or not. Considering that they outnumbered Waller's Foote at least 2:1 and that Waller's Horse was useless in the streets of Devizes, which had been barricaded, and even more useless against the castle where Hopton had retreated, Waller understandably was prepared to accept a delay in the hope the Royalists might surrender.

The first Royalist Regiment of Horse that appeared from Oxford bringing much needed supplies was ambushed and routed. Prospects looked good and the best explanation for what followed was that William Waller did not expect that such a strong force of Royalist Horse would turn up just the day after the talks had started. The Kings Leutenant of Horse, Wilmot, had put together a scratch force of all the cavalry he could get, about 2000 strong and had ridden to relieve Hopton. William Waller, upon learning that such a cavalry force was heading his way that day, marched to intercept them before the Cornish should take courage and, even worse, sally out and attempt to unite with the Oxford Horse.

Presumably in moving on to interept them before they came close to Devizes, Waller could not find any ground that offered a good defensive position. Not only that, one has to assume there was no time to prepare man-made defenses. In fact he probably would have felt obliged to draw the Royalists to battle, because if they just rode off and united with the Cornish Foote they would have become an impossibly strong force to defeat. The prevailing view is that Waller actually deployed on a high hill, Roundway Hill, some 700 foot high, but still on relatively open ground, no tree cover, no hedges or walls or any other cover for the infantry.

What is less easy to explain is the behaviour of the Parliament Horse. Hasselrig's regiment was on the left of the Foot and Waller's with at least another regiment (presumably Popham's as he was present in the battle) was on the right of the Foot. Most estimates are that there were 6 Parliamentarian Regiments of Horse of up to 2000 men in total, but which regiments were these? Only five are mentioned in the summer battles: Waller's, Haselrig's, Popham's, Thomas Essex's and Burghill's and they had born most of the weight of the battle at Lansdown for Parliament. Richard Atkyns, who fought in the battle, talks only of Haselrig's regiment on the right wing. Each cavalry wing would have faced superior numbers. They may have been outnumbered by as much as 2:1 or more, especially Haselrig's Horse that had been so heavily involved in Lansdown.

I have no proposals of what one could have done under the circumstances but Willliam Waller seemed to have set himself up for defeat which was what happened. One advantage the Parliament Horse had was that they were more rested than the Royalist Horse that had marched from Oxford. Whether there were 3 regiments or 6 regiments, they were probably of good quality, having had recent battle experience. Despite being outnumbered, they could have drawn out their front and charged full speed the Oxford Horse. The worst that would have happened was they would have been defeated but, if so, they would have had a better chance to escape.

What actually happened was that they stood where they were, perhaps hoping to frighten the Royalists and deter them from battle, or who knows what. Worse still, they formed a small front several ranks deep, whereas Popham's (presumably) regiment had formed not to the side but behind Waller's Regiment of Horse. So on both wings the front of the Parliament Horse was very narrow. This would have been ok for a charge on infantry but the Royalists had no infantry, apparently not even dragoons. The Parliament Horse did not charge anyway, they just stood there and allowed themselves to be charged on both sides by the much more numerous Royalist Horse. Horses are herd animals and no matter what their riders intend, they have an instinct to follow the direction of movement of other galloping horses. So whenever Horse charged standing Horse, the defenders were usually routed because the overall vector of movement induced by that charge is unfavourable to the defenders. Isaac Newton did not live then and Newtonian physics had not been discovered, neither people understood much about biology and animal behaviour, but in practice whenever there were horse battles the one who charged first generally routed the defender, so that ought to tell you something.

The Parliament Horse, finding itself surrounded by the greater numbers of Royalist horsemen, was put to rout on both sides without much of a fight. The situation was worst on the right wing. The Parliament Horse fled ahead of the Royalist Horse, unwittingly in the direction of a steep cliff. If you have 1500 horses galloping behind you, you would not even think of stopping, so the herd behaviour of the horses led that entire wing of Parliament Horse down the pecipice, many of them being killed. A few appeared to have escaped to the rear of the Parliament Foot, among them Alexander Popham, William Waller and presumably also John Locke, as he survived the war, though it is not known if he was at the battle. But the majority of the Parliament Horse had been routed. Some 1000-1500 Foote were now facing 2000 Horse plus about 3000-3500 Royalist Foote that came out of Devizes to participate in the battle. The odds were now hopeless. The Parliament army was completely destroyed and it was even lucky that all the generals survived, without even one of them being taken prisoner. Actually, Arthur Haselrig's Horse had been wounded and he was about to be taken prisoner when some of his men came to his aid and he managed somehow to escape in the confusion.

William Waller had gambled everything and had lost. Instead of a great triumph, there had been a crushing defeat. Now the Southwest of England was practically in the hands of the Royalists except for a few fortified towns. Bristol was taken by the Royalists, who now had at last a big city and harbour in the South of England. This was regarded as the high point of the English Civil War for the King.


-----------------------------------------------------------------------------------------------------------------------------
[MT2W FKoC AAR] Times full of Distemper
Reviewed by robinzx at the Critic's Quill, Issue 31
Medieval 2 Total War “For King or Country” mod: http://www.forkingorcountry.com/

[Geoffrey of Villehardouin]

The trouble with M2TW cavalry

I used to feel good about myself when I would run down a few infantry companies with my cavalry believing I had mastered a good tactic. That until I realised in some recent tests with the 1648 mod that cavalry is about 8 times as strong a unit as the same number of infantry. That means that cavalry shoots 8 times as much, has 8 times as high defense as the equivalent number for infantry and 8 times as high melee attack as the equivalent of infantry. Coupled with their charge bonus, that makes them into little short of supermen. While during the early Middle Ages heavily armoured cavalrymen were like tanks fighting poorly armed peasants, that might be ok. But during the Thirty Years War when infantry was either barricaded behind breastworks or fences (a bit as the Oda did at Nagashino) or behind pikes, attacking infantry with cavalry would have seemed about as foolish as eating soup with a fork. In 1643, Prince Rupert attempted to attack with his cavalry some poorly armed townsfolk barricaded behind some breastworks at the entrance of Birmingham and could not even get to them. It was only when he managed to get some of his men to fire at their flanks that the townsfolk decided to retreat back into town. Well, that was about the most successful cavalry action in the English Civil War against infantry. Charging at a stone wall or an earthen bank or a trench with cavalry could only gain some broken horse legs. Strictly speaking, infantry should not even be chargeable by cavalry at this period.

So take a cavalry unit, make it into 60 men, same stats as a similar infantry unit of 60 musketeers. Test them into a battle let’s say 60 horsemen vs 60 infantry musketeers. The infantry die 8 times as fast in the shootout. Change the mounted musketeers firepower to 1, keep the infantry to 8 – now they are killing each other at similar rates. You can do the same with the melee stats. Similar result. So in other words, to just balance the two in battle (forget the massive tactical superiority of cavalry commanded by the human player), you would need infantry with 8 times as high missile strength, melee attack and melee defense. That’s just to balance the strength of the units. Balancing the tactical ability of the AI to use its cavalry as well as the human player is a much bigger problem.

Here are some videos of cavalry with medieval M2TW stats (high charge, high melee strength, high missile strength) attacking infantry in the 1648 mod. First some late units with their original stats, as they appear in the game after patch 2.

That was not much of a contest ... You may say, ok, these are late units, they may do a little better, but the established doctrine at that time was to basically never attack infantry with cavalry, so why should they be able to win at all? They shouldn’t even be willing to attack, they should be routing once within musket shot range. Still, just to make my point, here is another example with early units, the Piccolomini Cuirassiers, with a morale (stat_mental in the EDU) of zero, backed up by early harquebusiers. Rather than the Cuirassiers routing in the first sign of musket shot, the infantry do ...

If you can trash infantry with the Piccolomini Curassiers, one of the weakest cavalry in the mod with stat_mental of zero, I don't know what need is there for infantry.


And here is a third example. Just two companies of early harquebusiers, supported by some mortars and some peasants. No Cuirassiers, no pike and shot units, attacking a stronger army, stronger in artillery and all its infantry being pike and shot units, including some of the best musketeers in the mod. Suicidal?

The human player wins with a convincing margin, losing only 111 peasants. The game engine is perfectly capable of turning the pike & shot period into a cuirassier & harquebusier period, even a peasant & harquebusier period, if you overlook the stats.

Basically, the only solution to this, since it is not always possible to nerf the cavalry 8 times, is to set stats of average cavalry and infantry to the same numbers, multiply infantry stats by 4 and divide cavalry stats by at least 2, to get a factor of 8 in their nominal stats – only to make cavalry as strong as infantry. Which is not even going far enough. Because most cavalry at the time would not even attempt to come close to musketeers.

Read on ...

[Geoffrey of Villehardouin]

More realistic cavalry for the pike & shot period

"If the Foote have possessed themselves of some place of advantage, as some wood, trench, or covert way, then the Horse are not to charge them, though equall or somewhat superior to them in number, in respect of such advantage."
John Cruso, “Military Instructions for the Cavallrie” (1644)

A custom battle with cavalry having nominally 8 times lower stats than infantry (missile and melee). In total 548 cavalry vs 560 infantry, an attack John Cruso would have advised against. Stat_mental for infantry was around 10-12. Stat_mental for cavalry was between 1-5 and charge was 5 for the Hussaria (which is the highest I have in the EDU), 4 for the Cuirassiers. These cavalry stats would still have allowed for an easy victory, had I not also minimised the amount of ammunition cavalry gets. Otherwise they would sit back and shoot the pikemen rather than charge them. Even so, the cavalry wins when historically it should have probably lost. The battle was at normal difficulty. Winning can be more difficult on the hard and very hard levels.

The inability of the M2TW engine to fight in formation is half the problem with M2TW battles. The other half is that the player can use what are essentially arcadish exploits, rather than historical tactics, to win battles. The computer player is handicapped against the human player in these two respects. Using high cavalry stats to win battles would seem a bit of an overkill, especially for the pike & shot period. Even nerfed 8 times compared to infantry, the cavalry is still stronger than infantry. And the musketeers are supposed to be barricaded behind breastworks not standing out in the open waiting to be run over by the cavalry.

[Geoffrey of Villehardouin]

As another test of cavalry, a custom battle between Saxony and Wuerttemberg from the 1648 mod, version 2 up to patch 2. Both sides have exactly identical units. Very hard battle difficulty with minor tweaks on the Xeryx XBAI and a few other files besides the EDU. As above, cavalry stats were nominally an eighth of the strength of infantry.

The Cornet Cuirassiers had a charge of 4, the Piccolomini Cuirassiers a charge of 2, with melee attack of 2 compared to the infantry musketeers having a melee defense of 16. Missile strength of the Harquebusiers was only 2 (arquebuses) and Cuirassiers had 8 (pistols) with much lower projectile accuracy than missile infantry, while the infantry musketeers had missile strengths around 18-30, similar to other mods from this period and to version 1 of the mod. Range was 250m for infantry missile units, under 60 for cavalry missile units. Despite all that, cavalry is clearly still stronger than infantry, and was the most decisive factor in winning the battle.

The strength of the BAI is in its use of infantry and especially the melee units. So with this unit balance the Saxon centre was crushed and the survivors had to retreat and skirmish until the AI pikemen were routed. From then on, the Saxon cavalry that had been kept safely out of trouble had a free rein. Final score: about three fifths of the human army was lost in the battle.

[Geoffrey of Villehardouin]

Shogun II Total War Shimazu campaign.

After realm divide, the once loyal Sagara clan have rebelled against their Shimazu overlords. Shimazu Yoshihiro put all who could be spared from the Kyushu island garrisons together into an army and moved swiftly to besiege the Sagara capital Higo before the Sagara could attack the undefended Shimazu hinterlands.

The only way to survive the Sagara cavalry attack appeared to be a defensive position on a hill with the rear completely protected. Once their cavalry was repulsed, the Sagara went into the defensive and set an ambush in a wooded area. As half their force was bow ashigaru, Yoshihiro simply ordered a frontal attack that sent the Sagara clan bowmen into a flight, leading to a complete rout of the rest of their army.

Had the Sagara won, the Shimazu campaign would have been in shambles since money woes had caused the disbandment of many units, ships were in disrepair, the provinces on the brink of revolt and an Oda army had managed to get past the Shimazu front and was ransacking the cities on the army's rear on SW Honshu.

[Geoffrey of Villehardouin]

Siege of Buzen - Ashigaru vengeance

One of the side effects of Realm Divide was the threat of naval invasions. An Oda fleet carrying an army in the Inland Sea was sunk but a Homa clan army disembarked on Kyushu. The castle of Buzen on a strategic position overlooking the Kanmo Straits was their likely target. It was only guarded by two companies of ashigaru and their retainers. Shimazu Toshihisa sailed from Suo with his bodyguard and a company of light cavalry while a company of ashigaru from Tsukushi also reinforced the castle. The invaders had a strong contingent of cavalry and a few companies of infantry, including some yari and bow samurai. Fortunately the AI keeps their cavalry back in sieges, which gave the defenders a chance of success.

The ashigaru excelled defending bravely against the first assault of the enemy infantry and also against the second assault of the dismounted cavalry. You can watch near the end of the movie a lowly bow ashigaru asking the enemy general to leave the castle and then winning the ensuing sword fight.

[Geoffrey of Villehardouin]

The Black Ship

One of the most troublesome aspects of Realm Divide surely is the threat of naval invasions. It turned out that as our army was outside Kyoto, the Ikko Ikki were sending an army by sea supported by the much feared Black Ship.

The Shimazu were fortunately controlling the western seas and the Shimazu heir, Shimazu Yoshihisa, also known by his nickname Hongo Kiyouji, was in charge of a largish fleet not far from Echiden. Time was of the essense and this was a third battle against all odds the Shimazu could simply not afford to lose.

A heroic naval battle follows, a death struggle, in which several of the Shimazu ships will go down while facing the biggest threat to Shimazu control of the western seas.

[Geoffrey of Villehardouin]

With the rear cleared from resistance and the sea routes secure, at last it seemed the time had come to bring the war to the Ashikaga Shogun. But you can never get too cozy in your sense of security. At that most inconvenient time, six pirate ships turned up near the trade nodes. Six spanking new medium bune ships, obviously bought with a loan from a foreign power that was not eyeing too fondly the growing Shimazu might.

There were enough military ships between the various trade nodes to take on the pirates but the danger was that might leave our trade ships undefended. So a small force sailed from our naval base at Bungo to become the core of a fleet that would hunt down the pirates. Cunningly, the pirates left the trade nodes alone and swiftly turned and attacked the force sailing up from Bungo.

A fire bomb kobaya was scouting ahead and was first to spot the pirates. It was not an easy fight and the Shimazu came close to losing some of the ships but in the end disaster was averted.

[Geoffrey of Villehardouin]

The days of the old Shogunate are over.

The Shimazu army marshal Tanegashima Tomokata along with the Shimazu daimyo's youngest son Ietsune have attacked the outposts of Kyoto.

Part 1: the sally

In response, the Ashikaga Shogun Yoshiharu sallied out with his entire army. A battle could decide the fate of the Shogunate.

The Ashikaga deployed in a forest on a hill. Scouts were sent forth by Tomokata but there are no signs of activity. But the Ashikaga had been seen entering the forest so they had to be there. The cavalry and some samurai were sent on the left to flank the hill under Shimazu Ietsune, while a small but battle-experienced force of samurai were sent off to flank the hill on the other side.

The main army under Tomokata advanced in the cloud dragon formation (as in Stephen Turnbull's Osprey book "Samurai Armies 1467-1649, rather than the exact formation used in game): archers in front, melee units echeloned further back in a looser formation, cavalry as bodyguards protecting the general and finally some naginata monks as the rearguard. That formation was recommended when the army was stronger but the enemy was hidden in difficult ground, just right for the occasion.

The bow cavalry on the left identified the Ashikaga Shogun and his retinue but their arrows were useless in the forest. The Samurai on the right advanced cautiously and came across some Ashikaga bow samurai, whom they charged. Immediately the entire Ashikaga army burst into life. The left wing of the Shimazu army with the cavalry joined the action immediately, with Ietsune charging the Ashikaga archers on that side of the hill. And the main army lunged forward, with the bow ashigaru providing some support, ineffective as it might be within the forest.

Placing their army in the forest made some sense, as long as the Ashikaga were to start using their bows while the Shimazu army was in the open. It seems the BAI has only a sketchy understanding of the tactical advantages of taking cover in a forest. Or maybe the Ashikaga were only hiding in fear? Perhaps fearing, indeed, the Shimazu might, the Ashikaga had camped right next to the map's edge, on the reverse side of the hill, so that the Shimazu infantry charged downhill into their ranks, having taken very few casualties from the Ashikaga archers. Once the bow samurais were routed, the entire Ashikaga army fled, leaving the generals behind to hold off the attackers. Three of the generals died in battle, with the Shogun only escaping after all his bodyguards had been slain. But enough men survived for the Ashikaga to still be able to put up a decent defense at the walls of Kyoto.

Part 2: the siege

Kyoto was besieged by the Shimazu daimyo Takahisa and the Shimazu army marshal Tanegashima Tomokata. Their combined armies surrounded the city. Tomokata led the main attack to draw the Ashikaga garrison to one side, while Shimazu Takahisa and Shimazu Ietsune attacked the walls on the other sides.

The gates had been sabotaged by our ninjas. For the Ashikaga, however, there was no way out.

There was at first a surprising spirited resistance, especially on the walls facing Tomokata and also on the western side, where Ietsune was leading a smaller force. However, the walls on the side of Takahisa, who turned up with a full army stack as reinforcements, were left completely undefended. Takahisa's army had led a victorious campaign on the northern side of western Honshu as Tomokata was advancing along the Seto Inland Sea coast. It was a strong and veteran army and once inside the city, the issue was decided. The Ashikaga Shogun Yoshiharu died a proper samurai's death - in battle. Ietsune was among the first to step onto the Kyoto citadel. With its fall, Kyoto fell and with it the Shogunate passed from the Ashikaga to the Shimazu and their daimyo Takahisa.

Victory! Glory! May the name of the Shimazu live on for a thousand years!

[Geoffrey of Villehardouin]

Rome II cinematic battle

Though Angry Joe has some good points (see his hilarious videoclip http://www.youtube.com/watch?v=P_QK-lcW8a8) over issues I actually have mostly not experienced yet myself, Rome II is an improvement over the original Rome in the graphics department. For some of us it matters, for me graphics put me off playing the original Rome. Animations have also been improved over Shogun 2, for example the routing animation where in all previous mods the unit about to rout stopped, most men turned around and began to walk away as if nothing was happening, before they started to actually run. There are also fewer examples of horses going through other horses (so far, fingers crossed). The landscapes are as good as Shogun 2 and the settlements are a vast improvement over Shogun 2. Athens, for example, is so carefully made it is breathtaking.

The BAI I have is reasonably good. It may have been patched since I only bought the game after the initial patching. About as good as Shogun 2, though different. Except that I sometimes win open campaign battles without a fight when the AI forgets it is supposed to be defending its flag. Although this battle was fun like other custom battles I tried, there are a couple of things you will see in this videoclip that are not quite right.

The Romans, once the chariots run into them, become tangled up in an amorphous mass and never manage to get back out into some kind of decent deployment. Of course that was my intention in using the chariots but maybe it should not be that extreme. The BAI guy needs to work out a way to get the Roman infantry at least into some kind of formation, they should not be just running here and there and coalescing into amorphous mobs during battles. Romans fought in maniples and the Greeks in phalanx formations and these followed some rules and have to follow some rules to be effective. A phalanx formation should not be flanked, there had to be a continuous line and effective coordination with the units on either side, so these formations were inflexible but very effective on the defense (like the Saxons in the historical battle of Hastings in M2).

AI controlled Roman maniples should have the ability to toggle through formations, archers in front or javelins in front or javelins in the second line and melee in front or spearmen in front, etc. To do that rapidly, the Romans kept gaps between the maniples. Although in Rome II Greek armies are deployed in phalanx formation and Roman armies in maniples correctly, once the battle starts the AI army formation goes out of the window. There must be a way to keep some of that (low tolerances in movement helped with the M2 BAI). One thing that seems to be gone is the guard command, where in fact that command might have had some potential in maintaining overall formation, if it could be used better in the game engine. Once this can be mastered, units and armies capable of fighting with more discipline (e.g. the Romans) can be made more expensive. There is too much predictability at the moment in the sense of how the BAI will respond well at the local level with relative disregard of the overall tactical situation - in the case of this battle forming a dense mob that was a perfect target for the Pontus artillery.

EDIT: Having started a campaign with Epirus with my only settlement with a port under blockade by a massive Roman fleet by turn 4 or so (plus being attacked from land by the Spartans) and the other settlement having endured maybe 6-7 sieges in the first 8 turns - on normal! - I can safely say this harder than any other campaign I have ever played. Whether it is due to the patching or the specific faction, unfavourable stars or what else I do not know, since I have not had the game in its original release. I went back to a previous save to deal with the Roman blockade, still the campaign ended in game turn 16 after Larissa had fallen to the upteenth siege. Apollonia was besieged by Athens and Ardiaei with three large armies plus a naval force when suffering from food shortage (I had no money to build anything except in the first 2-3 turns). The battle odds were in the region of 10:1 against Epirus. Simultaneously Pyrrhus was besieged in newly conquered Brundisium by the Romans with 17 ships (!) and an army outnumbering Pyrrhus by about 6:1 based on the game odds. My only spy had been assassinated and part of the Apollonia garrison had been poisoned in the previous turn by an enemy agent.

[Geoffrey of Villehardouin]

Yashima Taira short campaign on normal difficulty

The Yashima Taira is meant to be a hard campaign but given their position and their advantages (master shipbuilders, merchants and sailors) and their proximity to the trade nodes relative to the other playable factions, their campaign on the contrary seems one of the easiest. They are ideally suited in building a maritime powerbase in the Seto Inland Sea from where to expand and taken over the rest of Japan, or just the 25 provinces they need to win the short campaign. Most victory provinces are within that region except for Sagami in Eastern Japan that can wait until they are the dominant sea power and able to move by sea to Sagami and take it without upsetting too much the Minamoto.

In this campaign at normal difficulty, the army concentrated in capturing Okayama (Bizen) and Tsuyama (Mimasaka) both of clear military value as they can provide good quality or cheap units for the land army. Together with Akashi (Harima) they can form the nucleus for expansion in Honshu. By securing alliances with the Hasebe and Urakami on the northern coast of Western Honshu, a strong base was formed that would resist future attacks.

The Yashima Taira diplomat in the meantime headed for Kyushu, where he secured the allegiance of Kanoya (Osumi) and Kagoshima (Satsuma). From these bases our trade fleets could get additional support vessels and trade ships to quickly dominate the three trade nodes (incense and Chinese texts) on that side. The Junsatsushi next secured the allegiance of Saito (Hyuga) and the Taira forces that were subsequently raised in Kyushu took Fu'nai (Bungo) by force to serve as an additional naval base, starting a war with Hata, the most dominant faction in Western Japan. This is where the first videoclip picks up.

The Hata responded to the invasion of their provinces in Kyushu by rushly attacking Okayama. It was a clash of Titans but protected by the walls of Okayama the Taira had a distinct advantage. This is the first of three large battles in the first videoclip. In all of these the odds were equal or against the Taira. In the second battle, the Taira forces having taken Kokura (Buzen) in Kyushu turned against Fukuoka (Tsukushi) which they captured despite being outnumbered. A relief force arrived too late. In a battle near Kokura it was defeated and destroyed.

Kyushu soon fell completely under the rule of the Yashima Taira and Shikoku followed its fate soon afterwards. Inevitably Realm Divide led to war between the two Taira factions. Taira Tsunemori took Fukuhara after it had sworn allegiance and his son Tsunetoshi went a step further and took Kyoto by force. That act triggered a response from both the cousin Fukuhara Taira faction and surrounding clans such as the Kitabatake and Sasaki. Tsunetoshi had to endure two great sieges of Kyoto against massive odds. Its defenders fought almost to the last.

Kyoto had to be abandoned when it could no longer be defended but Tsunetoshi was a die-hard fighter flying the motto "War is Slaughter" and was determined for a comeback.

Placenames are different in ROTS compared to those in vanilla Shogun 2, which are shown for comparison next to them in parentheses.

[Geoffrey of Villehardouin]

Angry Joe and others have lamented the absence of a campaign AI in Rome II. This is Rome II post patch 2. I did not have the earlier version and do not know what the situation was really there. Nonetheless, a game is not a person and you can even break a person if you try hard enough, let alone a game. This videoclip shows you how hard can the Rome 2 AI be if you let the game take its natural course. But you can break it and I can do it too, and then the game becomes so easy to win it is simply a waste of time to play it.

These let's play-like sequences and battle scenes are from a Rome II Epirus campaign on normal difficulty. Unless you break the AI by capturing two settlements in the first 3-4 game turns, before it becomes active, this is the kind of AI behaviour you will have to contend with. Passive it is certainly not!

If you cannot beat the AI, maybe you can break it. But be fair to the game: if you do decide to break the AI, do not complain that there is no campaign AI behaviour in the game. And it would be an illusion to believe that when you play against nobody, not even against the AI, you are playing on hard difficulty, or on very hard, or on legendary, etc.

What probably happens is that once the human player becomes the most powerful faction, which is easy to do at the start of the campaign by quickly capturing just one or two settlements (since AI factions never seem to start with more than 2), then the CAI becomes passive - which normally ought to happen far into the campaign when the human player's faction should have 15 or 20 settlements not on game turn 4 with 4 settlements. Then to complain that there is no AI is pointless. That is in essence what a passive AI does - it does nothing. Just sits around its settlements hiring some units and waiting for the inevitable. The problem currently is that there is no middle ground, at least for the Epirus campaign: it is either near impossible to win even on easy difficulty, if the CAI is active, or else you can walk through the game without an opposition. Neither is much fun.

If someone has any hints on how to win the campaign once the CAI has become active, i.e. without the human player attacking before the AI attacks, then please post away.