So I think it would not be entirely impossible to mod long scenarios, particularly if the Gettysburg AI will itself already be better (which shall not mean it was bad in TC2M!) and it will be more modable. I am really looking forward to the new game!
Jan
Iy is possible to write long, long scenarios. The problem is the troops cannot stay together for a fight of that length. Morale, fatigue, and loses would force it out of action rather quickly. This is NOT how the battles were fought in the Civil War. Most fights would break down after a few hours.
In most cases I would agree with you Mark. However, July 1 is a battlefield where fresh troops continually enter into the focus of the town of Gettysburg The Union first corps and the eleventh corps, though separated, have their fates mingled depending on each to hold their own fronts--the fall of one will directly lead to the flanking of the other. To make this into short separated scenarios will disserve the ebb and flow of the first day's battles. The first day sets up beautifully for a grand multiplayer scenario with 4 players, each taking on a separate corps. The Confederates have short term objectives of breaking the enemy before them on the ridges and the long term objective of finishing that night on Cemetary Ridge. The Union's goal is to hold off the Confederates for as long as possible, even at the sacrafice of two corps. The scenario would not feel right unless it began at first contact and ended at nightfall of July 1.
I think MarkT's point here is that long scenarios are hard to do because units just don't hold up to sustained fighting, as they may have done during the war. I found this out while working on scenarios for Shiloh, in real life Sherman's division held for 4 hours against repeated Confederate attacks, but no matter what I did, either Sherman or the attacking Confederates would be completely routed after only 1 hour. The AI in TC2M, although very good, just doesn't understand the concept of falling back to regroup after the first attack stalls, instead it usually annihilates itself in one all-out attack. The only way I could offset that was to make the scenario shorter. Now, if Sherman's brigades and the attacking Confederate brigades were gradually introduced into the battle, rather than being thrown all together at once, then the fight could possibly be made to last for several hours. So, it's a bit of a trade-off, finding a middle ground between the capabilities of the game engine and the historical facts. Just my experience...an interesting discussion here.
I agree with both of you. There were very long battles but troops tended to be cycled through the worst parts of the battle.
TC2M for all it's greatness really fails in 2 tactical areas. First being the difficulty in fighting a delaying action thus giving up ground for time. Secondly, it does not assault, regroup and then assault again and again. I never got that feeling from the game like you do in the Gettysburg movie with Chamberlain on the left repelling multiple assaults.
This maybe needs a separate thread but I'll put it out there anyway. Every time a unit came under fire it's casualties reduced it's command and control ability and it's morale and resulted in the general fatigue and degradation of the unit. By falling back and regrouping that command structure could be regained and the unit morale improved as fatigue reduced. This allowed the unit then to form another attack. TC2M just does not do this very well.
I think that once a unit suffered about 30% casualties, it was pretty much done for that battle and was not used again for some time, (months), until it was brought back up to strength. This was more so for Union units. After Gettysburg, III Corp was out of action for months, (6 or so), due to the casualties it suffered on the 2nd day of the battle. Ultimately, III Corp never recovered and was dismantled in March of 1864 and the units reassigned to other Corps.
So I think that the shorter scenarios are kind of a natural break in the action that simulates a regrouping. The longer scenarios and long open play battles just don't do this and if you keep your guys rested/resupplied and the AI does not, then eventually the end is certain.
Regards,
GregB)
Last edited by Gfran64 on Wed Feb 04, 2009 8:00 pm, edited 1 time in total.
Now that I understand what you're talking about, I heavily agree with you. However, that is just the AI. If you tc a regiment and put a commander nearby, you can send it back into the fight.
I guess Gregg got that right, the short scenarios and carryover feature simulate those breaks, regrouping and moving on with new orders in some way. He is also right that TC2M does not do a good job with those kind of thinks, at least in stock.
In real battles, troops never fought for extended times, but with many brief or longer lulls in between. Sherman's Div in the Sunken road at Shiloh, for instance encountered something like 8 assaults in those 4h, with breaks in between (check the literature for more details, "Battles and Leaders" had a nice personal report on that as far as I remember). Also, if a division fought for x h, that sometimes means that one brigade/part of the div was fighting while another was idle for a while.
So, lulls were common, and continuous pressure as in modern day fighting very rare. Even at Wilderness, or Antietam (where Jacksons forces fought a series of short, isolated mini-engagements, like around the Dunker Church, and only few units (Hood two BG's were some) participated in more than one part of that fighting).
For the 3 day scenario I am figuring out, I tuned down pretty much all the casualty/moral/fatigue related parameters so that units can't be annihilated so quickly (though elite units can loose heavily as they stand longer), but fall back and retreat earlier with high fatigue. Indeed I see AI keep them out of action for 2-12h, depending on leader & unit quality etc. That nicely captures the essence of resting units, like Heth DIV on 2nd day, where he was just kept in reserve. After a while, those units come forward again (or defend VPs again, if Union). Just one annoying thing is that I have to "Aform" Brig_Lines and Divs in regular intervalls (lets say 6AM, 12AM, 6PM) so that AI regroups rallied units. Otherwise TC2M AI often just leaves them scattered somewhere (but fully recovered), I can't say why.
With the reduced loss numbers, I indeed get longer fire-fights, more fallbacks and maneuvering. With higher fatigue costs due to march and fighting, and much slower recovery, I tuned down the number of melees the AI would start. I get pretty historic behavior with that.
This way neither side is entirely bled down by the 3rd day, and total casualties close to historic levels. Also, units are never permanently in contact, unless the player behaves "unhistorically" and applies continuous pressure. However, due to the fatigue, even good players will find their limits really quickly because of tired units. Although I have been scripting the entire thing myself, and know exactly who the Confed are to do on day 1, I recently got whipped by AI for my first time -- ok, I had Reynolds, Bufford and Howard stick to their historic placements and tactics, but that should have been good enough expecting AI to attack piecemeal. Attacking piecemeal, it seems, is less problematic if fights take longer and casualties are lower -- that way AI will get more units into the fight before it is decided -- which likely means that it is not piecemeal any longer...
Side-note for those familiar with TC2M scripting:
Can I create recursive eventhandlers?
I.e. I call "evtblabla" at time 12:00, and it contains a line that calls itself at ~3600s? Would that work, or are eventhandlers deleted after first excution?
Side-note for those familiar with TC2M scripting:
Can I create recursive eventhandlers?
I.e. I call "evtblabla" at time 12:00, and it contains a line that calls itself at ~3600s? Would that work, or are eventhandlers deleted after first excution?
As far as I can tell, they are deleted/de-activated once they are triggered. But, you could use an "evtcont" and an entry in the "time var" to repeat a command after a certain amount of time.
Example
evtfighting>advance>time var "60"
evtcont>advance>time var "120"
Say this unit becomes engaged at 07:30:00, at 07:31:00 it would advance, then again at 07:32:00. But after that, the unit would not advance again while fighting, unless you or the AI told it to.
I'm very thankful for all you modders that know this stuff. I designed and wrote it all, but I could not have answered that question if my life was on the line. Thanks!
This would work for simple cases, but what I need involved generally all units of the AI controlled side/formation, so it would get very lengthy.
For instance I use the following EH 3 times a day (morning, noon, late afternoon) to improve AI tendency to reform BG's with scattered regiments (rallied) into close formations.
...
evtranBRIGLINECSA1 Jos_B_Kershaw Aform:Brig_Line
evtcont Paul_J_Semmes Aform:Brig_Line
evtcont William_Barksdale Aform:Brig_Line
evtcont William_T_Wofford Aform:Brig_Line
evtcont Robert_E_Lee ranevt:BRIGLINECSA:1 14400
But the last line won't call recursively anymore because the EH just executed and is deleted. So I could define 3x3= 9 times the same EHs (3 days), but with ~40-70 BGs on each side, this is a bit inconvenient.
It would be good to make EH's a little more capable, for example how about this? evtintrouble(delete=false,execute_between={10:00,12:00},random_delay=0,additional_events_for_start={objArmy1("Peach_Orchard","McLaws",3000);})
Attributes would depend on the EH type course.
Another thing are my "in trouble" handlers that somewhat prevent reckless attacks and also helps AI to retreat its Arty in order not to be overrun too quickly. AI unfortunately does not like to pull it back by itself, or relocate on occasion.
...
evtintrouble William_Mahone Aorders:Hold
evtintrouble David_Lang Aorders:Hold
evtintrouble Carnot_Posey Aorders:Hold
evtintrouble Battery111 Aretreat
evtintrouble Battery112 Aretreat
...
Here it would be easier to define a generic function, that could take the calling unit as an argument, such as...
evtintrouble William_Mahone Mygenericfunction
So if anyone has a good idea to script such events in a more efficient way, let me know!
In general, for scripting more complex (longer) scenarios, user variables, some way if generating hierarchic conditional test structures (if then else, case, for while...), new and more detailed eventhandlers and new functions(_pos=get_position "Battery112"; if (_pos select x = 1000 && _pos select z = 5000) then {zzzz}) would be awesome. This is probably the only thing I really miss in TC2M... Then I could create better workarounds and control the flow of the battle much better.
I hope some future version of War3D will include that...
Last edited by Janh on Fri Feb 06, 2009 11:09 am, edited 1 time in total.
