Unlike the other Battle Reports on this site, this particular battle is not an experiment to see how a particular principle or strategy works in practice. Instead, my son was depressed after a bad day at work and needed some cheering up. I did have two ideas that had been floating in my head for a while: I wanted to play with the new Siege Engines, and I was coming to believe that deploying blocking terrain near control points helped to inhibit the power of ranged units.
It has also been my experience that with blocking terrain properly deployed, custom defensive units like Titi Testudo and Wall of Steel could resist ranged attacks long enough to win the game on control terrain victory points. I thought if people saw this happen, it would help to illustrate my point. The joke is, I did have a unit that was able to withstand an unbelievable storm of ranged attacks, but it was the Ballista Invicti.
I didn't have days to prepare, so we used Nate's standard Kyudo-centric army with only minor modifications. The Way of the Bow rule variant chosen was one of two currently being considered for the new version of this ability.
If this unit has not previously moved or attacked this turn, its range dice score hits on ties.
The Wall of Steel was not only a good defensive unit, but it had the ability to heal the Ballista, which is a construct unit. I brought the Snipers because in past games Nate's Fireburst power was a major headache, and I was hoping the snipers could take the little flaming spirits out of commission. My plan was to take the control points with my defensive units, use my archers to slow him down in the middle, and use the Gauls to cripple his archers when they moved in to attack. Naturally, this plan did not survive the ambush turn.
Interestingly enough, I had intended to play with the Ballista Velox so I could get the extra movement speed, but I miscalculated the number of build points and had to use the Invictus instead. This would have a profound effect on the result.
We played a standard game with two central control terrain points covered by blocking and hindering terrain. This varies from the suggested demo setup in the rulebook mostly in that the blocking terrain is moved close to the control terrain so that melee units have a better chance of surviving once they get there.

Nate deployed the Eternals and the Rapid Fire unit on the left side of the battlefield. In the center, he placed his mounted cavalry, the Ravens, and one of his Kyudo controllers. On the right, he placed his second Kyudo controller behind a terra cotta screening unit.

I put the Titi Testudo on the right and a pair of Gauls on the left. In the middle, I had my Ballista and the Flamma Coniciens artillery unit, the Wall of Steel (in position to heal the Ballista), the Snipers and the third Gaul.

Nate won the dice-off and took the ambush turn. He is four for four on ambush turns. Honestly, I don't think I'd be able to cope if I went first.
Nate's first act was to push the Rapid Fire unit to get it onto the left control point. He carefully positioned the unit so that the blocking terrain prevented me from shooting it with my Ballista, thus proving my point about the importance of terrain placement in protecting melee units from archers.

Nate knew I would attack the Rapid fire will my Gallic sorties, so with his remaining two order points, he brought his left Kyudo Controller into position to kill the Gauls when they attacked.

This was my fourth game against Nate's Kyudo army, and I have a special hatred of the Rapid Fire unit: it's fast, it heals itself, and it has two Fireburst icons. There was no way I was going to let it live. I moved up both Gallic units. Using Hustle, these units can move 12 steps: I put one on the control terrain and another on the far side of the Rapid Fire unit where it was within striking distance of the Kyudo Controller. Moving the two sorties into position cost 4 order points.

I attacked with the first Gauls. Because the Gauls have Charger, attacking costs no order points. This Gallic unit still had Fervor, so the presence of the second Gaul on the other side gave me an extra attack die. The attack was 8 dice to 2, and I scored 5 hits, bringing the Rapid Fire down to two figures.

Nate knew I was going to attack again, but he reasoned that I have such bad luck with dice in the normal course of things that it was worth while to keep the healer.

The second Gaul didn't have Fervor, so I attacked with only 7 dice. Nate pointed out that if he'd rolled only a little better I would have failed to make the kill, and therefore keeping the healer was the right decision.

This, however, was only a theoretical consideration. The Rapid Fire unit had been killed, I was uncontested on the control point, and I had 2 victory points.
With my remaining 4 order points, I brought forward the Ballista and the Wall of Steel. I wanted to keep these two units together because the Wall of Steel can heal the Ballista. This would prove key to the Ballista's survival.

Nate decided to kill the Gauls. The Eternals were in position to attack both Gallic units. There was a direct shot to the Gauls poised to attack on the next turn, while the line of sight to the Gauls on the control point went through hindering terrain. Nate decided to shoot the control point Gauls first so he would have the benefit of Way of the Bow when firing with fewer dice. He rolled 6 hits, taking them down to only one figure.

The Eternals have a healer, so Nate can push them with impunity. He did so to attack the nearer Gauls. This time he rolled a full 8 dice and this time he also got six hits. Both Gauls had been reduced to a single figure.

Nate had used 4 order points. He didn't want to leave me on the control point, so he moved his mounted archers up and used them to finish off the control point Gauls. This unit normally shoots with 8 dice, but the hindering terrain reduced it to 7. This made little difference in the result: the Gauls had no defense and only one hit point.

The score was now 2 to 1 in my favor, but I had a nearly-dead unit in range of his left-side Kyudos, so it was still anybody's game. As it turned out, however, I had a strategic advantage that as yet neither of us fully understood.
The Ballista's Direct Fire ability could do 2 points of guaranteed damage at a distance of 8 steps (4 formation bases) provided I regrouped first. If I pushed, I could do a total of four damage, and then heal the push using the Artificer power of the Wall of Steel. The total effort would be 8 order points-- my entire turn.
I decided to go for it. There was no dice rolling involved, just move (2 order points), regroup (1 order point), Direct Fire (2 damage, 2 order points), push to Direct Fire again (2 more damage, 2 order points), then heal from the push (1 order point).

It's important to note that Direct Fire is a ranged combat order, not a special order. This means it costs two order points instead of one, and you can push to use it twice. It's also a factor that Direct Fire works from a greater distance than a normal ranged attack. Even though I'd just done four points of damage to Shanyu's Javelin, it was going to have to move in order to fire back.
Nate decided the Ballista was a danger and he would use what was left of the Javelin to kill it. He moved the unit up into range (2 order points) and fired, 8 dice to 6, doing 2 damage. I downed one roman legionary and injured a heavy legionary. Annoyed, Nate pushed to shoot again, this time rolling better and doing 4 damage. This downed my heavy legionary and I applied the remaining three hits to the ballista figure itself.
With his last two order points, Nate shot my almost-dead Gauls with the Eternals, using 8 dice to 0 with Way of the Bow for a total of 7 damage. The Gauls, and several generations of descendants, died, giving Nate another victory point and tying the game.
The Shanyu's Javelin, however, was down to half strength-- six hit points-- and the Ballista had taken what was effectively only a little scratch.
I began my turn by healing the Ballista for two hit points, then pushed the Flamma Coniciens to get it into range to shoot the Shanyu's Javelin. This left me with three order points.

Using two order points, I shot the Javelin with the Flamma Coniciens, using 6 dice with Fire Arrow. This did four damage.

With my final order point, I used Fireburst to kill the Javelin. This netted me two victory points, so I now led by 4 to 2.
For the first time ever, Nate wished he had more close combat units. The Ballista's Siege Defense ability gave it a small bonus against ranged attacks, but combined with the presence of multi-peg units on the base, this bonus was enough to keep it alive and relatively undamaged against a ranged assault. Without close-combat units to take out the Ballista's Direct Fire attack and punch past its special defenses, a new strategy was needed. Nate decided to move up both his Raven Fire unit and the Kyudo Controllers. He reasoned that one of these units would get hammered, but enough strength would be left so that he could double-team the Ballista and whittle it down to size. He then began plans to threaten the control points. On the left, he healed the Eternals to bring them back to full strength. On the right, he moved the terra cottas up into position, carefully measuring the angles so that the blocking terrain prevented either of my archers from shooting them.

I decided to use the Flamma Coniciens to try to take out the Kyudo Controller in the middle. I used two order points to move them into position and two more to shoot, 6 dice to 3 with Fire Arrow, yielding 4 hits.

A subsequent Fireburst attack did 3 more damage, and a Direct Fire attack from the Ballista added 2 more. I had consumed 7 order points and reduced the Kyudo Controller to three figures.

With my last order point I used Artificer to heal the Ballista back to full strength.
It was Nate's turn again. In order to use my Flamma Coniciens unit against the Kyudos I had brought it into range of the Ravens. Nate used one order point to Fireburst the Flamma Coniciens for two hit points and then shot it, 9 dice to 3. Incredibly, he only did two damage.

This was unacceptable, so Nate pushed to shoot again. This time, he got 7 hits, which was more than enough to kill the unit. I had lost half my artillery, and Nate had tied the score at 4 victory points.

Nate used two more order points to move the still-undamaged Kyudo Controller on the right forward. Since all my Fireburst units were dead, he allowed them to come into contact with the screening terra cottas he had moved up previously.
It was my turn again, and I pondered how to take out the Ravens. With the Ballista, you are always asking yourself if you should make a normal ranged attack (5 dice with this particular unit) or use Direct Fire. More often than not, however, the lower damage potential of Direct Fire takes a back seat to its incredible range. To shoot normally, I would have to spend two order points to move. The Direct Fire attack, however, could be done in place. I attacked the Ravens twice with Direct Fire, doing 4 damage, then used the Wall of Steel's Artificer ability to heal from the push. With my remaining 3 order points, I moved the Titi Testudo unit closer to the control terrain and moved my Sniper unit forward. With the Ravens so close, I felt I would need the Snipers to take out the Fireburst-throwing Goryo figures.

It was Nate's turn again. His central Kyudo unit had only three figures, and he would have had to push either of the others to get them into range for attacking the Ballista, so he decided to attack with the Ravens. He used 2 order points to move up the Ravens and 1 more to Fireburst for 2 damage. I removed one of the heavy legionaries, and then he shot 6 dice to 6, doing 2 damage.

I removed one of the formation legionaries and injured the other heavy legionary. Nate pushed to attack again, this time scoring 5 hits. To hit with every single die against an equal number of defense dice is an incredibly lucky result (3 would be normal). Ironically, the probabilities would reverse in just two more turns.

I downed the heavy legionary, then applied three hits to the ballista figure. Using the Ravens, Nate had managed to down all but one of my white-circle figures. The ballista was virtually defenseless, and it would take three turns to heal it back to full.
With his final order point Nate regrouped the Ravens for defense.
I used 1 order point to heal the two formation legionaries back onto the Ballista. I decided that from a purely statistical standpoint, I was better off shooting with a normal ranged attack instead of Direct Fire, so I regrouped the Ballista for its best range attack (5 dice). I shot the Ravens with the Ballista and got 2 damage, which I found really annoying.

Feeling I had little choice, I pushed to attack again, doing 2 more damage.

I'd used 6 order points and had two left. The Raven Fire unit now had only two Goryo figures left. To reduce their flaming ability, I moved up my Shanyu's Dirk sortie and used Sniper to kill a Goryo.

The Ravens were down to a single hit point, but Nate still had two completely undamaged artillery units, and they were both Kyudos. One was protected by a terra cotta screen. I had a half-dead Ballista, two defensive units, a sniper and a Gaul sortie that was trapped back in my deployment zone.
It was time once again to try kill the Ballista. Nate threw a Fireburst at it with his last Goryo (1 order point). I applied the damage to the last white-circle formation legionary. I didn't want to start injuring the ballista until I absolutely had to, because a serious attack was coming. That serious attack was from the Eternals on the left. Nate pushed them to get them into position to shoot and attacked with 8 dice to 2. The expected damage from this roll was 5, but incredibly, he did only 3, which I applied to the ballista figure. The Invictus had survived its near-death experience.

With his last order point Nate healed the Eternals from the push.
The Eternals were not only in a position to do a Way of the Bow strike on the Ballista, but they were in uncontested ownership of the left control point. I had, fortunately, a weapon in reserve I could use to take them out: the Gauls. To get the Gauls into position to attack the Eternals, I had to get the Wall of Steel out of the way, so I moved it in front of the Ballista (after using Artificer to heal the Ballista a little).

The Gauls could now move directly forward and, by pushing to move again with Hustle, slide over into the Eternals, snapping into position for an attack.

The attack was 7 dice to 3, yielding 6 damage. This was one of the few situations where the Intuition ability gained from having Savros Kuznetzov on the Wall of Steel made a difference. (Of course, his Artificer ability had already paid for itself a dozen times over.)

To keep the Eternals from healing, I used the Shanyu's Dirk to Sniper the Yuezhi Healer.
The Eternals were now in a bad way. The other artillery unit was all the way on the other side of the field, stuck behind a screening unit of terra cottas. Now, that screening unit of terra cottas was also keeping the Titi Testudo on the same side of the field at bay, so it had a useful purpose, but the Eternals could not get help, and if they attempted to escape the Gauls, the free breakaway attack would all but kill them.
With my last two order points, I Direct Fire the Ravens, killing them and gaining two victory points.
Though the game seemed like it had been a bloody mess to this point, in fact there had been a lot more injuring than killing. I had killed one sortie and two formations; Nate had killed two sorties and one formation. We had played 14 turns, and the score had been tied for most of that. This had changed, because I was in the lead, 6 to 4.
Nate now moved the terra cottas and the right side Kyudo Controller up near the control point. The terrain prevented him from moving freely, and he was unable to shoot the Ballista, but he could fire at the Wall of Steel.

The expected damage in this situation is around 1. There is a 55% chance of doing no damage at all. Instead, Nate rolled a critical hit, and I rolled garbage. He scored 7 hits, an event with a probability of only 2%.

Because Savros is so important to the survival of the Ballista, Nate decided it was worth his while to push to shoot again. This time the results were a little closer to the normal range, and he got 3 hits. In a single turn, the Wall of Steel-- a unit with 11 defense dice against ranged attacks-- was reduced to only one hit point.

For the last time, I used Artificer to heal the Ballista. Nate was in a position to take the right control point, so I needed the left one, which meant using the Gauls to kill the Eternals. The Gauls attacked 7 dice to 0. It took two attacks to finish the Eternals, The first attack did 3 damage and the second did 5. The Gauls were down to six figures, but they owned the left control point and I was up to 8 victory points.
Nate still had a crippled Kyudo Controller in the middle, and he had a nearly full-strength one on the right with a screening terra cotta. Each of us had a control point, and he would almost certainly get an extra 3 for killing the Wall of Steel on his next turn. It was therefore necessary to start crippling the Kyudo on the right so I could safely bring the Titi Testudo into position to contest the right control point. I pushed the Ballista to Direct Fire the Kyudos twice, downing four more figures.

I wasn't going to be able to heal the Ballista any more, so I opted to down the green-circle figure when I pushed, sacrificing movement instead of defense. I needed to keep the Ballista alive long enough to cripple the right side Kyudo, so defense was important, and with the Ballista's incredible range, movement was not an issue this late in the game.
Nate started his turn with 3 victory points for the right control terrain, making the score 8 to 7. I was still ahead by one point.
Rather than go after Savros, Nate moved his right Kyudo Controller into position to shoot the Ballista, feeling that if he could finally kill the thing, he could still win.

The first attack did three damage.

The second attack did no damage.

With his last two order points, Nate moved his middle Kyudo unit into position to be able to shoot the Gauls and Savros.
At the beginning of my turn, I gained 3 victory points from the left control terrain, making the score 11 to 7 in my favor.
I still had Savros sitting on the Wall of Steel, so I healed the Ballista again. I used two Direct Fire attacks to do four more damage to the Kyudos on the right, bringing them down to 2 figures. I had three order points left, so I used two to bring the Titi Testudo closer to the right control point and one to move the Shanyu's Dirk into position to do a Sniper attack on the Kyudos.
Nate now got himself 3 victory points, closing the gap to 1 point in my favor.
The middle Kyudo Controller could now shoot the Wall of Steel without moving. Though it had only 3 attack dice, the Wall of Steel was defenseless; Savros was killed instantly, netting 3 victory points and giving Nate a victory point lead of 13 to 11.
Nate now shot the Ballista with the right Kyudo. Due to its decrepit condition, it had only 2 range dice against 6 defense dice, but Way of the Bow applied, so it did 1 damage. This was applied to the Ballista figure. Nate moved the middle Kyudos into position so they could shoot the Ballista next turn, then regrouped the terra cottas for attack instead of movement, expecting an assault from the Titi Testudo.
I took 3 victory points, making the score 14 to 13. I then Direct Fire attacked the middle Kyudos, taking them down to one figure (2 order points). With the Shanyu's Dirk, I used Sniper to take the right Kyudos down to one figure (1 order point). I had five order points left. I based the terra cottas with the Titi Testudo, positioning them so they were on the control terrain. A close attack on the terra cottas dealt 4 damage and I took 1 back damage.

Using Armored Rush, I took down two more terra cotta figures, leaving the unit with one four hit points. Nate had a maximum of 4 attack dice, and the Titi Testudo had 8 defense dice. After 20 turns, the game was effectively over.
While I had done a theoretical analysis of the Roman Ballista, I had never seen one in action before, and the result was stunning. The Ballista is a credible threat to any artillery unit that tries to attack it and is capable of recovering from damage. The Invictus model is particularly well-defended. Of course, a certain amount of luck was involved. In pure probabilistic terms, the Wall of Steel should have been the unit that kept holding on, while the Ballista was slowly ground into the dust. Nonetheless, the Wall of Steel can't punish units that get into range, and the Ballista can make them suffer mightily. The Ballista's extreme shooting range is also a big factor, enabling me to force Nate to come to me instead of the other way around.
It's also the case I was facing an almost pure artillery force. A more balanced army would almost certainly have a better chance against the Ballista. That in itself, however, is a significant development in the evolution of the game.
I feel that blocking terrain placed near control points is critical in keeping artillery under control. Unfortunately, it helped protect Nate from my artillery much more than it protected me from his. The unit that enabled me to sneak around the blocking terrain to get him was the Hammer of Bellonas. Indeed, one of my biggest problems in the game was having my last Hammer trapped behind the Wall of Steel unable to engage my opponent. That was a major pain.
The alternative version of Way of the Bow made positioning a key factor. Here again, the Ballista had an advantage. My normal artillery unit died almost instantly once I brought it into play, because a half-dead Kyudo could still crush it. The Ballista, however, forces opposing artillery units to move before they shoot, and that kept the Kyudos from doing their usual dirty magic.
One game is not enough to prove anything, merely demonstrate the possibilities; however, it is clear there are lots of possibilities in a Roman Steam Ballista. We have no alternate stat figures, we have fewer special abilities, and our ranged figures suck, but now we have possibilities.
This could be seriously fun.