Lew
Private
Posts: 29
Joined: Fri Jan 09, 2009 4:03 am

Artillery Analysis (what type to buy, and why)

Sat Jan 24, 2009 11:11 pm

Answering the question "What type of artillery to buy?" involves some calculation. This spreadsheet [edit] was an attempt to answer that question.

NOTE: This spreadsheet is been removed as it was too inaccurate to keep up and I lack the time to fix it. Thank you to all who critiqued it - much appreciated!

User avatar
lodilefty
Posts: 7616
Joined: Sat Aug 11, 2007 3:27 pm
Location: Finger Lakes, NY GMT -5 US Eastern

Sun Jan 25, 2009 12:13 am

How well does this line up with:
http://www.ageod-forum.com/showthread.php?t=12717

?
Always ask yourself: "Am I part of the Solution?" If you aren't, then you are part of the Problem!
[CENTER][/CENTER]
[CENTER]Visit AGEWiki - your increasingly comprehensive source for information about AGE games[/CENTER]

[CENTER]Rules for new members[/CENTER]
[CENTER]Forum Rules[/CENTER]

[CENTER]Help desk: support@slitherine.co.uk[/CENTER]

User avatar
Major Tom
Lieutenant Colonel
Posts: 275
Joined: Mon Jan 12, 2009 4:00 pm
Location: Alexandria, Virginia

Mon Jan 26, 2009 3:41 pm

Thanks, Lew. Excellent work, and thanks for posting the spreadsheet so it's viewable in the message.

This is a nice supplement to Dixicrat's work, and his doesn't cover USA at all.

User avatar
lodilefty
Posts: 7616
Joined: Sat Aug 11, 2007 3:27 pm
Location: Finger Lakes, NY GMT -5 US Eastern

Mon Jan 26, 2009 4:19 pm

Major Tom wrote:Thanks, Lew. Excellent work, and thanks for posting the spreadsheet so it's viewable in the message.

This is a nice supplement to Dixicrat's work, [color="Red"]and his doesn't cover USA at all[/color].



Actually, it does ;)

his example on how to use it is CSA, but all the USA data and analysis is there....
Always ask yourself: "Am I part of the Solution?" If you aren't, then you are part of the Problem!
[CENTER][/CENTER]

[CENTER]Visit AGEWiki - your increasingly comprehensive source for information about AGE games[/CENTER]



[CENTER]Rules for new members[/CENTER]

[CENTER]Forum Rules[/CENTER]



[CENTER]Help desk: support@slitherine.co.uk[/CENTER]

User avatar
Major Tom
Lieutenant Colonel
Posts: 275
Joined: Mon Jan 12, 2009 4:00 pm
Location: Alexandria, Virginia

Mon Jan 26, 2009 4:32 pm

lodilefty wrote:Actually, it does ;)

his example on how to use it is CSA, but all the USA data and analysis is there....


Ah, I did not realize that! I haven't been able to look at his new 2.0 version yet. His original data covered only CSA.

Coregonas
AGEod Guard of Honor
Posts: 1072
Joined: Sun Nov 25, 2007 9:34 pm
Location: Barcelona-Catalunya

Tue Jan 27, 2009 11:59 am

I propose you to recalculate all your EXCEL considering 2 aspects, as told in Dixicrat thread.

1.- Ignore MONEY for your analysis. Just count WAR SUPPLY costs, the real limiting factor.

2.-Consider another line with 6 lbs as 12 lbs.

These upgrade at a fast ratio in the long term campaign. By march 62, 40% of all 6 lbs are 12 lbs. By April 62, 60%. By June 75%... This is for CSA. For USA it is even faster.

conclusions dont change too much, but some surprises arise.

User avatar
denisonh
Captain
Posts: 196
Joined: Tue Apr 24, 2007 12:22 pm
Location: Northern Virginia

Tue Jan 27, 2009 4:04 pm

Coregonas wrote:I propose you to recalculate all your EXCEL considering 2 aspects, as told in Dixicrat thread.

1.- Ignore MONEY for your analysis. Just count WAR SUPPLY costs, the real limiting factor.

2.-Consider another line with 6 lbs as 12 lbs.

These upgrade at a fast ratio in the long term campaign. By march 62, 40% of all 6 lbs are 12 lbs. By April 62, 60%. By June 75%... This is for CSA. For USA it is even faster.

conclusions dont change too much, but some surprises arise.


I would submit for the Union money is the limiting factor rather than WS, so it is relevant to the USA on $.

Coregonas
AGEod Guard of Honor
Posts: 1072
Joined: Sun Nov 25, 2007 9:34 pm
Location: Barcelona-Catalunya

Wed Jan 28, 2009 12:41 pm

:cool: sorry I didnt know USA had any limits other than manpower.

User avatar
Dixicrat
General
Posts: 523
Joined: Wed Aug 13, 2008 8:55 pm
Location: East Tennessee
Contact: ICQ

Wed Jan 28, 2009 7:55 pm

Hi, Lew!

Two significant ideas regarding AACW Artillery have recently been developed on these forums. You should know about these, because they're far from obvious.

Here's the first.
For a long time, many people had believed that there was a limit of four artillery per division. Common knowledge said that more than that was a waste. Not long ago, it was discovered that this was a myth, and that the limitations of the number of artillery pieces which could be effectively used in a division was defined by what is called "frontage", rather than any hard-coded number, in the game code. Even so, there were some common misconceptions about what those ideas implied. Some people (myself included) were initially ecstatic at the idea of large numbers of powerful artillery in Divisions.

Jabberwock provided research in the thread, "Frontage", which actually provided the exact numbers of "how many, of what" could be effectively used, given the particulars of terrain and weather. Major Tom followed this with a thread of his own about Frontage, going into much greater depth. If anything, the seemingly inescapable conclusion of the data presented seemed to be that much, much more artillery could be used than had been previously imagined. I decided it would be a good thing to consider what the limits might be. I presented the pros and cons of divisional artillery in the thread, "Artillery Divisions". Based upon that research, I privately came to the conclusion that artillery in a division should number between five and seven, with eight or more being the rare exception.

The second significant "discovery" was that there are essentially two artillery ranges of engagement: "Long Range", where a weapon's range is of consequence; and "Short Range", where a weapon's maximum range is of no consequence. The surprising part was in the details: "Long Range" engagements only occur in clear terrain, in fair weather. (Clear Terrain being defined as clear/desert/prairie.) With such a variety of weather and terrain as the game provides, it seems that Long Range engagements are the exception, rather than the rule.

Each unit in the game is based upon what is called a "model". Those models are defined in the database which can be found in the modding sub-forums. Our forum's current discussion of Artillery is centered (mostly) on two ideas.

The first involves the models of Columbiads and Rodmans. Currently, there's a consensus that the current models should either have movement penalties (because weapons of such power would be heavier and slower), or be redefined as something else that's more historic. For the Rodman, it has been proposed that the model be redefined as a "3 inch ordnance rifle". This suggestion is controversial. [Edit 5 Jun 09: After reviewing the "columbiad/rodman" thread, I realize that this was your suggestion. I think your arguments were good ones, and some of your considerations have been used by our AACW beta team in the process of redefining and re-evaluating the current Rodman model. :) ]

The second involves adding new artillery units. For example, one proposal is for a 24 lb howitzer. This is mildly controversial: on one hand, some believe that this weapon is already included in the "siege weapons" battery; a different line of thinking believes that this would be a useful and versatile addition to the AACW arsenal.

Anyway, we've learned a lot, but there's a lot that's yet to be learned. And regarding your artillery analysis... I'd like to be among the first to commend your effort, and encourage you to do more! :)

I'll discuss the particulars of your spreadsheet in a different post.

User avatar
Dixicrat
General
Posts: 523
Joined: Wed Aug 13, 2008 8:55 pm
Location: East Tennessee
Contact: ICQ

Mon Jun 01, 2009 9:17 pm

Here are a few of my thoughts about your spreadsheet, Lew.
My experience of learning this game has shown me that what seems obvious at one point may later turn out to be something that's filled with implications you didn't see. For example, it came as something of a shock when I realized that artillery doesn't participate in offensive assaults. The only time that assault values are used are during defense. Another thing that's easy to miss is that assault operations have their own values. Thus, the Offensive and Defensive hit probabilities aren't used. Instead, assault "to hit" values are actually based on a 1-10 value; each "point" is multiplied by 5 to determine the percentile probability of an assault hit. For example, a 12 lbr with a value of 7 will have a 35% base chance of making a hit during each round of assault.

For this reason alone, you may want to re-evaluate many of your equations in the spreadsheet, since you've included assault data in many of your equations.
Another thing which seems to be missing from your analysis is a consideration of "cohesion". Cohesion is a significant factor in unit battle effectiveness. I strongly disagree with your conclusion that "Parrotts should be avoided". Firstly, the Federal 20 lb Parrott has the highest "ranged cohesion damage" value of all ordnance. Secondly, its range is greater than any other FA weapon, excluding Rodmans and Columbiads. As for the 10 lb Parrott, it also has a range advantage over smoothbore weapons such as the 12 lbr and 6 lbr. Consequently, at range 6, smoothbore weapons cannot return fire against the 10 lb Parrott. Finally, both Parrotts have superior Offensive fire to that of smoothbores. So: the combination of greater range and greater hit probabilities for Offensive operations is a strong recommendation for Parrotts, in my opinion. In my own work, I have yet to develop any constants that I'm satisfied with, and so I'm rather curious about your use of what appear to be subjective constants as coefficients. Why did you use them, and how are they derived? In particular: how have you derived the "Assault damage Coefficient" in row 17? I don't understand either parameter. I'm also curious about the "Replacement Coefficient" on line 26.Next:your statement that "6 lbrs are a waste of money and space until they upgrade to 12 lbrs" isn't something that I agree with. Six pounders are useful for at least three purposes: 1) building forts 2) providing "trans 4" entrenchment bonuses until "something better" can replace them 3) providing supplementary artillery for higher siege bonuses, when you just need a few more points. As but one example, a 12 lbr by itself has no siege bonus; but add a 6 lbr., and a siege bonus of "1" is achieved; and add a second, third, and even fourth 12 lbr, and that "useless" 6 lbr has become a catalyst, enabling each new 12 lbr to contribute to siege value. Finally: if you'll provide me the means, I'll be glad to send you an equation that I think you'll find very useful: the "Power" equation, otherwise known as "Combat Efficiency". This equation (or actually series of equations) is used by the AGE engine to calculate the "power" value that is displayed on the unit counters. It's very useful for designing new models, and also learning the relationships of the various properties of each model.

Lew
Private
Posts: 29
Joined: Fri Jan 09, 2009 4:03 am

Fri Jun 12, 2009 7:01 am

Thank you very much for your well-considered response. I hope that other readers of this thread have learnt as much from it as I have... :)

...or I'll get whupped pretty hard in MP! :bonk:


Dixicrat wrote:Finally: if you'll provide me the means, I'll be glad to send you an equation that I think you'll find very useful: the "Power" equation, otherwise known as "Combat Efficiency". This equation (or actually series of equations) is used by the AGE engine to calculate the "power" value that is displayed on the unit counters. It's very useful for designing new models, and also learning the relationships of the various properties of each model.
Gladly! PM sent.


Assault values: Absolutely agreed and I entirely take your point.

The 20lb Parrott's higher cohesion attack: Is not modeled at all in my spreadsheet, and therefore I withdraw all claims about it.

The 10lb Parrott's greater range and greater offensive hit probabilities: I propose that the spreadsheet does (or, with coefficients suitably tweaked, could) deal with these advantages and judge whether they are sufficient to merit a cost so much greater than that of 12lb smoothbores. I'm pretty sure they're not.

The coefficients: All of these come out of my ass (well, not quite, but almost) and are intended to be adjusted by the user depending on his own operational situation.

The Assault damage coefficient is 66.7%, which means that the spreadsheet thinks that a given point of damage or cohesion loss is 2/3rds as valuable if inflicted in assault as it would have been had it been inflicted in the previous ranged combat rounds. Is this correct? Haven't a clue. Is it a reasonable approximation? I (currently, subject to correction) believe it is.

The Replacement coefficient is incorrect (I figured this one out on my own ;) ). At 0.5, the spreadsheet thinks that the cost of replacing an artillery unit is half as important as the cost of building it. The reason for this value was that I figured that 1/2 of all guns might reasonably need replacing over time, which to me still doesnt' seem a bad guess. However, each replacement chit on average replaces TWO units, which means this value should instead be 0.25.

Excellent info about 6lbers! Thanks for it!

Another point: Columbiads and Rodmans now move at a speed of 70% of that of other artillery. The spreadsheet's claims about their worth are now obsolete.


I've deleted the spreadsheet. Had it been a little off, It would have been worth keeping. Had I more time, I'd take your and others' critiques and revise it. As it is, better to make it disappear.

Return to “AACW Strategy discussions”

Who is online

Users browsing this forum: No registered users and 11 guests