Jump to content
Game-Labs Forum
Sign in to follow this  
Aetius

Some Information on Scaling

Recommended Posts

If the most important thing is wiping out enemy brigades, then wouldn't reducing the enemy size to make wiping out enemy brigades easier be a good thing?  Assuming you are interested in manipulating scaling of course.

Share this post


Link to post
Share on other sites

I don't think it works that way because you are dealing with the entire Confederacy and not just a particular army. So the smaller the force you are fighting, the smaller the effect is on the overall impact to the confederacy as a whole. So if you are fighting against a reduced force, your efforts don't impact the overall army sufficiently to drop their levels of training and use of advanced weapons. If you want to have a real impact, I do believe you actually want the enemy to bring as many troops as possible to the minor battles.

Of course, forces from different theaters can be interjected and are and that can impact training and armory and it appears that at certain times the army resets a bit. So, for example, although it would appear that I have the Rebs on the ropes after the Supply Raid (1 star infantry armed with re-bored farmers and no star skirmishers), at Nansemond River their skirmishers and infantry bounces back to 2 stars with 3 batteries of 3 stars. And after having wiped them out, at the Siege of Suffolk the Confederate army is packed with 3 star infantry units, though the artillery and cavalry is still at 2 stars. However, even when I wipe them out at the Siege of Suffolk, they get a big bounce in men, training and armory in the intelligence report. But when I then fight them at Chancellorsville their cavalry is almost non-existant because I have been slaughtering their cavalry for quite some time and even when I bring a full cavalry division of 5 brigades of 300 cavalrymen to the battle. Their general quantity of artillery is low and 2 star and their infantry is also I believe 2 star as well. Though I have to tell you those results were had when I decided to take my army after Nansemond River and without reinforcing (except adding my cavalry division before Chancellorsville) then played through the Siege of Suffolk and Chancellorsville at double time. Officially, I am now at the Battle of Suffolk for actual game play.

So I believe that while you can actually influence the overall quality of their army by wiping out as many brigades as possible, there is also some scripting in there that allows them to bounce back at certain points (3 star artillery batteries at Nansemond River and 3 star infantry at the Siege of Suffolk) and you have to start again to reduce their quality. This stands to reason as, if not, you wouldn't have a very challenging experience in the overall campaign.

Share this post


Link to post
Share on other sites

I absolutely agree with your last statement about the scripting. I'm more skeptical of your theory about the impact of wiping out a smaller vs larger force is accurate, but I only have my own anecdotal data to work off of. Unfortunately getting actual data to confirm one way or the other involves a lot of time consuming testing.

My current thought would be to use first bull run as a starting point. Fight the battle mostly as normal but with the goal of leaving as many enemy units unshattered as possible and clumped together. Make a save then finish the battle with all units left unshattered, half of enemy units shattered, and all units shattered. Then compare weapons and training as well as the actual unit sizes(using a recon 10 mod) to see if there is any noticeable difference. Enemy casualties should be fairly close on all 3 tests so if the unit destruction is important then hopefully this should show major differences.

Performing a more extended test across multiple battles or a set of side battles and major battle would probably give a better set of data, but would be considerably more time intensive. The types of reinforcements are definitely semi-random if not completely random so repeat testing across a series of battles would be difficult. I'll see if I can get to the bull run test sometime soon.

Share this post


Link to post
Share on other sites

Sounds interesting, however, because the gaining of experience/efficiency is a gradual process, I'm not sure if you will see much distinction. If there is any you should see it in the next minor battle of River Crossing.

I've got an idea I will try and we can compare results.

Share this post


Link to post
Share on other sites

So this isn't the greatest test, but I do have the results so it's better than nothing. Taking the battle of Phillipi and comparing the post battle AI numbers at Bull Run. The biggest impact on AI Size is the randomized reinforcement amounts.

1) I face the normal 9560 men at Phillipi on Legendary. I kill 5716 and capture 158. Reinforcements are 11500 recruits. 

Intelligence Report: 28-33k Training: 37-42 Weapons: 8-13. Size at bull run 31988 men 47 guns.

 

2) I face a modded 24056 men on Legendary. I kill 13772. Reinforcements are 11000 Veterans and 12100 recruits.

Intelligence Report: 31-36k Training: 45-50 Weapons: 14-19. Size at bull run 36555 men 58 guns.

 

3) I face a modded 23255 men on Legendary. I kill 12001. Reinforcements are supplies and 14400 recruits.

Intelligence Report: 25-30k Training: 38-43 Weapons: 14-19. Size at bull run 33188 men 52 guns.

 

4) I face a modded 23255 men on Legendary. I kill 13503. Reinforcements are a transfer of 21200 recruits.

Intelligence Report: 30-35k Training: 40-45 Weapons: 6-11. Size at bull run 38789 men 62 guns.

 

If it weren't for test case 3 I would have concluded that increasing the size at Phillipi directly increased the AI pool even though I killed more men. But when comparing 1 and 3 it's clear that these numbers are, at least at this point in the campaign fairly predetermined. A better test case would clearly be testing differing kill rates against the same total at a battle, but after more than a dozen playthroughs of Phillipi testing the mod I might go insane if I have to play it again anytime soon.

Share this post


Link to post
Share on other sites

I can also confirm at this point that scaling does work mostly the way I have described in the past. The algorithm behind it is significantly more complex, but the basics are essentially accurate. If anyone wants to actually discuss the algorithm feel free to message me but I won't be posting the code here directly for a variety of reasons.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×