[IMPORTANT] Can Union AI hold Cemetery Hill on first day?

[Nick Thomadis]

With latest patch version 0.93 there have been various changes to AI. One of them should improve Union AI defense tactics in all personalities. These tactics should be effective enough so that AI holds Cemetery Hill and Culp's Hill with bigger probability. Please let us know your experience as you play vs Union with any AI personality of your choice.

The task of Union AI is very hard, considering that it plays even (with no gameplay cheats) and is vastly outnumbered by Human Confederate forces that come from West, North and East. We hope the Union AI manages to hold its ground valiantly. Write please your feedback. It will help a lot to improve this particular AI issue.

 

You can reply here or in Steam forum:

http://steamcommunity.com/app/306660/discussions/0/616189106402661247/

[RichardMcCor]

Nick, also help the CSA to win on day 1 against you as Union. I played 3 times yesterday with no clicks for the union against the CSA (determined) and it lost each time. Union got minor victory. I want determined to win easily when i do nothing as union.

[Nick Thomadis]

Nick, also help the CSA to win on day 1 against you as Union. I played 3 times yesterday with no clicks for the union against the CSA (determined) and it lost each time. Union got minor victory. I want determined to win easily when i do nothing as union.

Yes Richard, already done for next patch

[RichardMcCor]

ah yes, that classic bit of code will do it

 

if player = richardmccor

then 

    set csa_attack_strength = 500; set union_defense_strength=0;

    ignore_richardmccor_mouse_commands(25% of the time);

    union_arty_ammo=dud;

    csa_arty_ammo=WWIILevelExplosives;

done

[long5hot]

No click battles, the Union is holding the Cemetery and Culps easily on 1st and 2nd days versus the AI playing Confederate (doesn't matter which setting). The AI is vastly improved over what it was, but still not enterprising enough to force the issue.  The no-click Union wins easily.

Will try now playing as Confed.

[long5hot]

Ok, Confed vs Cunning, attacking the Cemetery and Culps on day 1. Here's an album describing the battle in detail...

http://imgur.com/a/1Nls5#0
 

Things the AI did well, showing significant improvement on previous versions:-

• Holding units in a line
• Reacting to an enfilading situation by telling the threatened brigade to fall back
• Wheeling brigades to face new threats
• Committing reserves

Where the AI did poorly:-

• Sending brigades into pointless isolated attacks with little if anything to gain.  Especially brigades covering key objectives like Culp's Hill, leaving them wide open and forcing reserves to be sent to help out.
• Not managing Skirmishers and Videttes well at all.  When they should have been a pestering nuisance harassing me from behind my lines once bypassed, they just stood around with no idea what to do and were hence a non-factor
• Not seeing a broader enveloping move coming until it was way way way too late

The AI made several key tactical mistakes with the Iron Brigade charge & the non-handling of Skirmishers, and two key strategic ones by not reinforcing its left flank and leaving its center too far forward where it could be isolated and crushed.  Indeed, the AI seemed to think it could advance on its center and its right & just ignore what happened on its left, and this in the context of a battle where it expected to be outnumbered and attacked from multiple flanks.  In the center this advance also made its artillery powerless to support its line.

 

In the face of overwhelming numbers, but with reinforcements expected, the AI would have been better served by trying to keep its line shorter, falling back where needed in order to do so while holding the objectives, and keeping attacking forays to an absolute minimum.  It also needs to be aware of what's happening on its flanks as its still very susceptible to indirect maneuvering.

 

Can Cunning hold the Cemetery on day 1?  I think not.  As this is essentially a defensive battle which it lost by being overly aggressive/ambitious and thus losing its ability to hold the objectives, I'd imagine that Dynamic and Determined will be easier still to defeat.

[Husserl]

We've done some further changes and we expect a better defensive behavior with the next patch.

[Harnis]

just did the same battle as you did long5hot, my outcome as Confederates against Cunning Union, I took Cemetery Hill, but they held Culp's Hill and Cemetery Ridge. Only reason why they lost Cemetery Hill was because they focused a large number of soldiers in the actual town of gettysburg rather than just slightly south on the hill. My attacks on Culp's Hill got nowhere, my attacks on the defending force in Gettysburg made very little progress, and I took Cemetery Hill because I flanked it from the south with 3rd Corp. Pettigrew, Archer, Davis, and Brockenbrough took Cemetery Hill and flanked city, while Lane, Scales, and Perrin faced south defending against the Union reinforcements. Those reinforcements were able to take and hold Cemetery Ridge before I could advance on it.

 

If the Union had centered its main defensive circle on Cemetery Hill rather than in the town of Gettysburg, It is doubtful as to whether I could have taken it. As it was, the outcome was a draw. It was a hard-fought battle and some of those Union brigades were very hard to uproot from their positions.

[long5hot]

Hmm, interesting that the outcome was different - did you follow the exact same strategy I did?  I guess I should have pointed out in my pics that I transferred quite a few brigades to the right wing (the south) in the initial stages, so I only fought a holding action in the town - the whole plan was based around that enveloping move and that's why I had the troops to take the Ridge quickly, well before Union reinforcements arrived.  Culps would have been much harder to take for me had the Iron Brigade held firm & not abandoned it with its charge.

 

I'll play through it again and see if Cunning handles things differently.

[long5hot]

Ok, second play-through.  The AI handled things differently, but the end result was the same - in fact it was worse for the Union with a Stalingrad-style encirclement of its center.

 

Same problems as noted previously - its too aggressive and ambitious in the north which makes it try to hold a line that's longer than its capable of sustaining, draws its troops away from its objectives and makes it incapable of responding in strength to flanking maneuvers.  Late in the battle it again could have taken the Ridge if it had pushed, but despite having superior numbers of fresh troops it just stood back as if expecting me to vacate the area of my own accord.

 

Here's the album:  http://imgur.com/a/dbBGO#0

[Harnis]

Haven't had a chance to play it again yet.

 

But i don't know if I allocated quite as many soldiers to the south flank. All I did was pretty much just march all the brigades forward in the positions they were already in. No major flank relocations. But I would definitely agree with your conclusion that the Union AI puts too much effort into defending the north, specifically the town. Like I said, my battle was only a draw because of the reinforcements that stopped me from taking the ridge objective. But the Union main defensive ring was focused too far north in the town. If it had been focused on the actual Cemetary Hill, they probably would have won, because since the latest patch, Union brigades actually can hold their ground. And yeah they didn't make too many of those foolish charges like the iron brigade did in your battle. In mine, iron brigade and ames brigade did a solid job of defending culps hill the entire battle against two confederate elite brigades from Ewell's corps

[Nick Thomadis]

Thank you all for the reports. They help a lot. Can you play again with latest update 0.94?

[Harnis]

Just played the same battle again. Confederates Approach to Attack Cemetery Hill. Playing as Confederates vs Cunning Union.

 

I used the exact same strategy as last time. Sent Lane, Scales, and Perrin south to take the Cemetery Ridge objective. Sent Davis, Archer, Pettigrew, and Brockenbrough up Cemetery Hill from the west side of it, in line with Lane, Scales, and Perrin. I had Ramseur and Daniel on the Seminary Ridge hill, just west of Gettysburg. I had Doles, Iverson, and O'Neal advance south through the town of Gettysburg. I had Ewell's 1st Division advance from Benner's Hill to attack Culp's Hill from the north-east.

 

I'll start by mentioning this: the game crashed right at the end. I was taking Steam screenshots (ctrl-alt-f12), and when I went to take the second one, the game froze, stopped responding, and had to be closed. But the battle was literally about to end in the next moments, so I still got solid results.

 

Hill's 3rd Division (Lane, Scales, Perrin) advanced easily up the slope to the Cemetery Ridge objective, and were able to take it relatively easily. The arrival of XII corp and Stannard's brigade prevented me from using them to flank anywhere else though, and they were then stuck defending Cemetery Ridge from the attacking Union soldiers.

 

Hill's 2nd Division (Archer, Davis, Pettigrew, Brockenbrough) were able to slowly push the Union soldiers back up Cemetery Hill. They were able to reach the crest of the hill, but were not able to reach the objective.

 

Ewell's 3rd Division (Ramseur, Daniel, Doles, O'Neal, Iverson) did not make much progress. Over the entire battle, they were able to slowly advance and resulted in taking the town of Gettysburg, but they were not able get anywhere close to the Cemetery Hill objective.

 

Ewell's 1st Divison (Gordon, Tigers, Avery, Smith) had a rough fight. They made absolutely zero progress against the Union brigades holding Culp's Hill. The Union general actually sent two brigades to flank Benner's Hill, causing me to need to relocate 3 brigades from Ewell's 3rd in Gettysburg to circle around and stop them on the east side of Benner's Hill. Note that once they flanked Benner's Hill, killed some artillery units, they just sat there on the east side of Benner's Hill. This was possibly due to their damaged condition and morale, but it just seemed like a waste of a flanking maneuver. Also note that this was an aggressive move, relocating two brigades from Cemetery Hill to attack on the Culp's Hill flank that was in no danger of falling, meanwhile I was taking Cemetery Ridge. Those two brigades would have been much better to relocate south to the ridge to defend it rather than making a useless flanking maneuver. Then again, hindsight is everything, and no general is ever perfect, so I can tolerate some mistakes as long as they are "intelligent" mistakes (meaning they were intelligent actions that were carried out well, they were just the wrong decision).

 

The end result was me holding Cemetery Ridge, and the Union holding Cemetery Hill and Culp's Hill. Not sure what the casualties were, but they seemed pretty equal. So I'd say a minor (possibly major) victory for the Union.

 

Observations:

 

Union artillery on Cemetery Hill was firing more, probably due to better line of sight for them in this patch. Confederate artillery from Seminary Ridge was essentially useless, however. They had absolutely no line of sight to anything going on near the objectives. I had to move them all the way up on top of Cemetery Hill to start getting shots on stuff (mostly on XII Corp reinforcements)

 

The Union defensive position was centered on Culp's Hill and Cemetery Hill. They did not waste brigades piled up into the town anymore, instead they actually defended the objective. There was a point about one third through the battle where I just zoomed out, took a broad look at the battlefield, and thought "well shit.......how the hell am I going to crack THAT defense....". It was actually a nice feeling. I felt the challenge.

 

Melee combat definitely causes more casualties, but not too many. When I sent 3 brigades to stop the Union flanking Benner's Hill, they charged me and a melee ensued. The Union brigade broke, and took massive casualties for being a routing unit among two enemy brigades in melee with them. It seems like a good balance. Melee has more effect than before, but it's not over the top. Also, those flanking Union brigades easily mopped up two artillery units I had on Benner's Hill with a charge. So melee balancing seems much better.

 

 

Overall, awesome patch. This game is getting better with every update, and the AI (which was amazing from the start) is only improving.

 

edit: regarding the crash. Just started the same battle once again, took one screenshot, it was fine, pressed ctrl-alt-f12 to take another screenshot, and the game froze, exact same as before. Not sure how I can give more info about what happened. Is there a log file you'd want me to send?

[long5hot]

My Cemetery Hill playthrough vs Cunning with the latest update: http://imgur.com/a/GivW9#0

 

In summary, the strategy I used in my previous battles still works - I secured all three objectives for an Epic victory plus inflicting twice the number of casualties on the Union army.  Despite this, the AI was significantly improved. Things it did well:-

• It held its line extremely well from one end of the front to the other.
• It used its skirmishers to very good effect.
• It was far less prone to rash attacks that weakened its line or its hold on objectives.  The Iron Brigade did not charge off Culps, for example.
• It did not press too far north of the Cemetery.
• It was extremely good at falling back under pressure, maintaining the line and minimizing routs.

Things that need work:-

• It still did not recognize the danger posed by the increasing bend in its line due to my south flanking attack.  If it had pulled back further from the town earlier on, it would have had a shorter & straighter line to defend plus freeing up extra brigades to help counter my flank attacks.
• It only held one brigade (Bucktail) in reserve, and its deployment was wasteful having no impact on the battle (it charged at Sharpshooters well away from any objective).
• When its center was in extreme danger it still did not withdraw, preferring to try to fight on in an untenable position.
• As a result, the AI did not counter my enveloping moves on either flank and this lost it the battle plus a ton of men.
• Stannard was stuck on the edge of the hill where it enters the map and had minimal effect on the battle.  It did not like my occupying the Ridge before it arrived.  I tend to think this particular battle would be well served by extending the map a little further south.

To be honest the flank attacks are the way to win as Confederate, and to repel them the Union would have to detect what's happening and respond in a very quick & agile manner.  I don't know that any real-life general would have been capable of this, as from the outset of the battle the Union Army is in a very difficult position.  Otherwise the AI managed its army quite well. 

 

On a battle with no pre-existing kinks in the line to exploit and no weak flanks to turn I imagine the AI would be very tough to crack.  To this end, I'll next test playing Confederate on July 2 - Afternoon, All-out attack at Seminary Ridge.

[RichardMcCor]

Played single player as union tonight . Set CSA AI to different levels and played it on day 1 a.m. scenario 1  x 7 times.

Each time as Union i did nothing. No clicks at all. Just watched and hoped to see the different AI tactics unfold.

 

The CSA AI did best against Union in Offensive and Risky Mode - it got a draw only but captured Oak Hill and Ridge while doing it. Good use of his General if Offensive mode by the way.

The CSA AI as Determined also got a draw but did not actually capture anything more than Herrs ridge. Same with Cunning

All the rest lost to the Union with a minor defeat (minor win to me).

 

Offensive really gave it a good shot and flanked the union position to the north very nicely.

 

One thing I did notice was long charges. Very long charges. From McP Ridge to Oak Ridge by Petigrew  against Paul multiple times. Same result each time, i.e. ran out of puff before he got close, stopped, got shot up, ran, got shot up again while running. 

 

I sent in a few bug reports on the CSA ARTY who seemed to not be able to get into a good firing position sometimes and that hurt the AI's attack plans.

[long5hot]

As promised, here's my playthrough of July 2 - Afternoon, All-out attack at Seminary Ridge versus Cunning: http://imgur.com/a/uJSlP#0.

 

I chose this as the Union is defending a straight line with both flanks anchored in good defensive positions (unlike the Cemetery day 1 scenario), with plentiful reserves to manage and with armies of approximately the same size.

 

I'm sorry to say the AI did a lousy job.  Did not reinforce objective points, did not pull troops back from positions where they'd be surrounded, did not use Generals well, did not position Artillery effectively, did not maintain a coherent line, and left its reserves untouched while its frontline troops were being ripped apart.  As a result I could defeat a very strong Union army piecemeal and seize all the objectives.

[long5hot]

... and I replayed the Cemetery scenario against Determined. It repeated all the mistakes that prior versions of the game made with Cunning, with a similar outcome.

http://imgur.com/a/nMchO#0

[Harnis]

here's my recent playthrough of Confederates Approach to Attack Cemetery Hill, playing as Confederates against Cunning Union AI

 

http://imgur.com/a/JT5As#0

 

The Union again focused its defenses mostly on Culp's Hill and Cemetery Hill. Very light defenses on Cemetery Ridge in the south.

 

The Union AI didn't play badly, it just did not respond much to the dual flanking maneuvers by Ewell's 1st Division in the east and Pender's Division in the south-west. It had a solid defensive setup, and held Culp's Hill extremely well, and it only lost Cemetery Hill because it conceded the ridge too easily, and I was able to flank Union brigades there from the south and encircle them.

 

The Union consistently defends Culp's Hill extremely well, making use of falling back, counter charges, flanking, etc. They even had artillery support on Power's Hill in the south that was hitting Smith's Brigade when I attempted to send him around to flank Culp's Hill from the south.

 

This patch, the AI defends Cemetery Hill much better. It focuses forces on the Hill rather than in the town. The only issue is that the AI does not respond well to flanking attacks to the south on Cemetery Ridge. I mean, based on the initial deployments of the armies, the Union position is fine, but when you start to move brigades south to push up the ridge from the west, the AI seems to ignore it and just carry on defending the two hills north. With Culp's Hill so solidly defended, the AI could have afforded to move some reinforcements to the south. As soon as the XII Corp arrives, the Confederates are basically stuck at whatever position they've reached at that point, and can't advance any further. So all it takes is a few more brigades on Cemetery Ridge to delay the flanking maneuver, and then as soon as XII Corp arrives, the flank will be secured, and that means the Cemetery Hill objective will not be threatened from behind anymore.

 

So I guess the conclusion, and the over-arching weakness in the AI is its ability to respond to broad flanking maneuvers.

[Nick Thomadis]

Really helpful your responses.

Next update will make the AI to counter flanking maneuvers more successfully.

[David Fair]

Buried with work & travel for the past 2 months - but I had time today to test the current version playing Union with AI on Cunning.

 

The AI is improved - I finally got to noon on Day 3 before the CSA ran out of infantry.  Previously this always happened by Day 2 no later than noon.

 

Union VP's = 51,000             CSA VP's   =   7,000

 

Union casualties = 19,651    CSA casualties   = 33,170

 

I notice the artillery loaded with canister and positioned on "hold" will not fire unless the target unit is selected.  Charging infantry runs up to the guns with no casualties and destroys the battery.  I'd expect to see the artillery fire one round before melee or the guns retreat.

 

The Corps jumping positions between phases still seems like an odd feature of this game.  

In the morning phase Culp's Hill is defended by XI Corps and Cemetery Hill is defended by I Corps.  

In the afternoon phase the positions of these formations has switched.

Not clear to my why Corps are baited and switched between phases.

 

The "Tilt-Shift Effect" produced a black band at the top 1/3 and bottom 1/3 of the screen - bug reports sent on the two games I started with this feature.  When I turned off the Tilt shift things worked just fine.

 

The massive artillery formations really lend themselves to panzer tactics.  Just set these babies on canister and move around the field like Philip of Macdedon smashing everything in their path with his phalanx at Chaeronea.

 

It just seems like there should be a better way to implement these artillery units.  

They reign supreme on the UGG battlefield.  

I put two of these artillery formations side by side and smashed the center of the CSA with these.  

Felt more like the Somme than the ACW.

I'd suggest toning these formations way down to something like historical levels to prevent the Panzer/Phalanx tactics. 

[Harnis]

artillery now take tons of casualties from melee, but I think they still take too few casualties from being shot at. I mean, a brigade of 2000 soldiers can fire a volley at a section of 50 artillery men, and only kill 1-2, but fire at another brigade and kill 10-15. Now, maybe we can assume that artillery units are in cover better, more spread out and don't present the same large mass of men that a brigade does, or whatever. I'm not sure. But it does feel a little strange sometimes.

 

I know the game is meant to be simple, with a focus on tactics rather than micro-managing unit formations and special abilities and whatever. But I think artillery does need some kind of limber/unlimber mechanic. Also, I think an entrenchment factor needs to be added in. Something like the longer a unit stays in a position without moving (rotation would be ok?), the cover rating for that section of the map increases to simulate them starting to dig in and make trenches, breastworks, barricades, etc. This would also work for arty. So if a battery is sitting on a hill for hours, they would be able to withstand multiple volleys from enemy infantry, whereas if they were caught in the open or shortly after relocating, they would be butchered.

 

It could be as simple as allowing artillery to rotate freely, but they need to limber in order to move. Limbering takes a little bit of time, but then movement speed is increased compared to now, so it would even be faster than infantry. If arty is caught while limbered, they would take massive casualties, kind of like how a unit caught in column formation takes more casualties. They'd need to unlimber before firing, which again takes a short period of time.

 

I think just the simple limber/unlimber mechanic would totally eliminate panzer tactics with artillery. And heck, if they don't want to introduce that level of micro-management (understandable), make it automatic, like infantry using column formation. Just make it that when you give them a move order, they automatically limber, and then unlimber at the destination. They could even make it so that extremely short movement orders would be accomplished by rolling the guns around very slowly.

 

Add to that the cover/entrenchment system and it would make arty behave more properly.

[long5hot]

I agree with everything you said Harnis, but add to limber/unlimber the speed a horse-drawn artillery should have but also force it to cross streams at bridges. That would make things very interesting indeed.

If anything just feels wrong about the game right now its how artillery behave.

[David Fair]

long5shot,

I'd agree that the artillery behavior is wrong - but for a whole list of historical reasons.

 

Rather than making stuff up it seems like using the data from Gettysburg and the ACW might offer some novel insight for an ACW artillery implementation.

 

Also - by putting artillery in its historical context the micromanagement of artillery could be vastly reduced by deploying it by brigade rather than battery.

 

 

You might want to google some images of the creeks at Gettysburg - in June you can walk/jump across most of these without getting your feet wet.  They weren't much of an obstacle in most places.

 

Also, recognize that the pace of artillery in UGG is glacial for the Horse and Musket era - and the ACW in particular.  

 

To cripple the artillery further with crossings at streams you will really insert a number of factors into the artillery arm that didn't exist at Gettysburg.

 

 

I'd agree that the artillery behavior is wrong - but for a whole list of historical reasons.

 

Rather than making stuff up it seems like using the data from Gettysburg and the ACW might offer some novel insight.

 

The most casualties any battery took at Gettysburg was about 35%.  

Artillery casualties on both sides at Gettysburg overall were 10%.

Why?  Because ACW artillery was agile and could (and did) limber to the rear at a trot or gallop to escape almost all close combat during the ACW.  

This agility is why you see so many guns captured on the Napoleonic battlefields and so few on ACW battlefields.

 

The CSA lost 35% of their front line artillery horses at Gettysburg - altering CSA artillery tactics both during Gettysburg and for the remainder of the ACW.

The artillery from both sides was in tact and in defensive positions all along the lines on Day 4.

Only 6 Union batteries were overrun at Gettysburg and the CSA only managed to capture and use the guns from Smith's Battery.

The Union retook used all but (I think 6 total) of the other guns in the batteries that had been overrun.

The CSA abandoned only 1 gun on the field at Gettysburg - the barrel had burst.

 

Artillery during the ACW inflicted only about 6% to 8% of the casualties during the ACW.

General Gibbon and Hunt compare the firepower of an artillery battery with a 200 man infantry regiment.

 

Why is artillery in UGG as effective as an infantry division?

Why is it so easy to wipe out batteries in UGG?

Artillery could, and did move, faster than infantry in many battlefield situations at Gettysburg - why is it so slow in UGG?

Why can artillery function in UGG like a Panzer division?

 

Rather than focus on the limber/unlimber question why not introduce the concept of "Horse" in the Horse and Musket Era?

Guns didn't attack because their horses would be slaughtered.

This immobilized the guns.

The loss of horses was much more fundamental to the ACW artillery performance than limber/unlimber mechanics.

The potential loss to horses is why artillery was primarily a defensive weapon during the ACW.

On the attack artillery was usually deployed on a ridge supporting the action.

If the ridge was carried then the artillery was moved forward to this ridge.

The process was then repeated.

Look at both the Peach Orchard and Longstreet's Charge on Day 3.  This is exactly what happened.

This is the same behavior ubiquitously during the ACW.

 

 

But as usual - argue what you think would enhance the game.

 

Let me know if you want me to repost the artillery statistics from Gettysburg regarding artillery.

 

 

This game has so much potential it is a shame to discard/disregard ACW data/fact in favor of fiction.

 

If you get the facts aligned with reality the tactics in will follow.  

This is a fundamental truth of all weapons systems.

It is also a consideration that might positively influence game design.

[Harnis]

But as usual - argue what you think would enhance the game.

 

Let me know if you want me to repost the artillery statistics from Gettysburg regarding artillery.

 

 

This game has so much potential it is a shame to discard/disregard ACW data/fact in favor of fiction.

 

If you get the facts aligned with reality the tactics in will follow.  

This is a fundamental truth of all weapons systems.

It is also a consideration that might positively influence game design.

hey man, easy, we're all on the same side here in agreeing that artillery does not behave properly right now, both for gameplay reasons AND for historical reasons.

 

I like your idea of deploying artillery by brigade rather than by battery. It does seem wrong when you're fighting a battle and you have 6 infantry brigades as units, but then you have 12 artillery units comprised of 3-5 guns.

 

I think we should try and narrow down the changes to be made to arty to a concise list. I'll give it a shot:

 

1) deploy by brigade rather than battery

2) limber/unlimber ability. Results in much faster movement when limbered.

3) higher vulnerability to infantry fire. Much higher.

4) further balancing of artillery damage to get closer to historical kill rates for artillery. Possibly balance this by increased damage to condition, fatigue, morale, etc.

 

Anything you think needs to be added to that list? Removed? Changed?

[long5hot]

But as usual - argue what you think would enhance the game.

David, your comment seems to have a personal edge to it, I'm not sure why. I'm all in favour of an accurate approach (rather than the totally fictional artillery we have now) and I happily concede that your knowledge of the history is far deeper than mine. If I inadvertently suggest something that's wrong then I'm happy to be corrected - case in point, thanks for pointing out the passability of the streams in the area.

Harnis, re vulnerability to infantry fire, I agree but that would also have to work in with your former suggestion about "digging in". As most artillery units wouldn't be moved that frequently, particularly for the army that's defending, they'd be far more protected than a unit caught out in the open. The horses too would be kept safe behind the lines until there's a need to limber, so they'd only be in danger if the artillery came under fire/melee in a limbered state.

Lastly if artillery is closely supported by infantry, the attacking infantry wouldn't have the luxury (as they do in UGG) of concentrating their fire on artillery crews & ignoring the enemy infantry, plus if the battery's in serious danger of being overrun the crews can run back from their guns behind the shelter of the infantry.

Also agree with the comments regarding the casualties inflicted by artillery historically. As I understand it, they did however have a potent psychological effect both on nearby friendly troops and enemies. Infantry did not like coming under artillery fire - those cannons inspired fear.