Brett's Games

The Roman War Machine

Fourteen Foes to Conquer

CIV IV
Beyond The Sword

Game Save & Replay Files
017-roman-war-machine-game-saves.zip

Computer Generated Logs
Courtesy of the BUG Mod
017-roman-war-machine-logfile.txt

In Game Notes
017-roman-war-machine-notes.txt

Crab Cake Cove
A Rich Seafood Start

Five seafood with forest and hills, make this the best start in my opinion in the first 25 rolls, others are free to disagree
Home Sweet Home

The Settings

Noble
Augustus Caesar of Rome

Map: Tectonics
Size: Small
Era: Ancient
Speed: Marathon
Landmass: Lakes 30% Water
Aridity: Wet

No Barbarians
Aggressive AI
Permanent Alliances
No Tribal Villages
No Random Events

All Victories

15 Total Civilizations
14 AI Opponents

Twelve of the AI Opponents are pre-selected and listed out in 017-roman-war-machine-notes.txt. My intent was to select the most warlike. But on this, individual opinions will differ.

Beyond the Sword 3.19
BUG Mod 4.5 [Build 2221]
BULL 1.4 [Build 243]
BAT Mod 4.1

The Pre-Game Strategic Huddle

I think a quick point counterpoint format will work best... for me, at least, if no one else.
A few other thoughts: Oh, I should, also, mention that I am enabling Normal Technology Trading, as any other setting is difficult on the AI. Now, I question whether normal Trading is fair to the Human. But it's been a while since I've played that way, so time to be annoyed and watch as the AI Collective advances.


And that's about all the ideas I have for the Pre-Game Huddle.

Let's start this sucker!


...spoilers...
...spoilers...

...spoilers...
...spoilers...
...spoilers...

...spoilers...
...spoilers...

Iron Poor

Plan On It!

I got Iron and that is good enough, fifteen tiles away from my capital, and it took me ninety odd turns to connect, but I got it, the image shows my first three cities, the iron source being down in the lower left, past the second city

In the above image, my capitol (Crab Cake Cove) is hidden behind the five seafood (in the upper right). Down in the lower left, is the sole available metal (Iron or Copper) and is a requirement for putting the Praetorian War Machine into motion.

Waiting until I discovered Iron allowed me to settle my Cities in such a way that not only was Iron connected, but the other City (Two Fish) makes sense in its supporting role.

Hey, I mean, I think this is the best settlement pattern. But without the Iron, I may have settled Two Fish three tiles north, so as to grab the Gold (and Stone, which would be a long time in coming).

Thus, I said I would wait until Iron to settle my Cities. And I am glad that I did.

Sweet Victory

My Victory was Total and Complete
745AD
T449

Score: 2,726
Normalized: 161,715

4 Civilizations Destroyed
7 Civilizations Vassalled

I believe that is my highest score yet (on a non-rigged map, anyway).
I should analyse that a bit further.

Build Analysis

Buildings Built
ItemCostBuiltHammers
Stonehenge3601360
Moai Statues7501750
Barracks150121,800
Courthouse360103,600
Forum45041,800
Lighthouse1804720
Library270102,700
Temple2402480
Monastery1802360
TOTALS4612,570


Military Units
ItemCostBuiltHammers
Praetorian90958,550
Longbow100212,100
Catapult100292,900
War 'Phant12091,080
Hvy Footman1402280
TOTALS15615,210

Thus, I spent about as many hammers on infrastructure as I did on military.

Maybe some commentary would be in order:
Typically, I play a weak early game, build up my Technologies, and steam roll the opposition at Infantry. So, there is a time transfer going on there: late game power in exchange for early game weakness. Many of the buildings have a similar trade off: an early game waste of hammers for more something (hammers, research, culture, gold) in the late game. Put simply, if the game lasts long enough (a million-zillion turns), then every Building is a wise investment. For the 449 Turns this game lasted, the math is a bit trickier.

And now, for a recap on the Unit Builds.
Next game (and yes, more and more as I write up the latest game, I am looking forward to the next game), I intend to play Augustus Caesar of Rome on a Standard Map at Marathon Speed all standard rules in play on the Monarch Difficulty Level, as a proof of concept. Is the Roman War Machine a viable strategy? Or did I just adjust the settings to meet my needs?

I think I will win. And as such, part of my strategy will be to milk my adversaries better. By this I mean: I believe I will be able to eliminate all the rivals on the continent on which I start (handily enough) with a Praetorian Rush (or Offensive, take your pick). So, I might as well play with my opponents, capturing Religions, Wonders, Workers, Unit Experience, Improved Cities, and so on. In other words, since Victory is assured, I need not rush into it. Let my victims shower me with Tribute.

Also, it bears mentioning, that there will be two sets of wars (or so I believe) in the next game, one with the neighbours on my home continent and those who I will require Optics to meet and Astronomy to attack. This game has shown that the initial Home Front War should be quick and painless. Thus, part of my overall strategy will be in transferring power from the early game to the late game (which is what I normally do).

But then, that is next game.

Graphs

What do they tell us?
a standard exponential population growth curve lasting 450 odd turns on marathon speed, reaching a total population of under 300

I played this game on Marathon Speed, which takes three times as long (1,500 Turns to reach 2050AD). So, at Turn 450 (or 449 to be exact), we are about a third of the way there. More importantly, the Total Population is under 300, which is not exactly high.

This was a War Game.
yellow gold, blue research, pink culture, research is always offset by a gold deficit, the end game culture rush is not nearly as prominent as in other games

This game did not last that long nor was Research ever my strong suit. Research (blue) was always offset by Negative Spending (gold), which I earned by capturing cities.

Next game, I wish to conquer my lands at a more measured pace (playing with my victims if that's what it takes) to keep research at a better level.

At the end of the game (as I often do), I switched from Research to Culture, hoping to grab those few extra tiles (Percent of Land Controlled) required for Victory. Up until that point, my Culture was virtually non-existent. Next game, I wish to follow a similar trajectory, as most Homeland Culture is wasted.

Combat Graphs

Maybe, we'll learn something new?
Every combat that I engaged in was recorded, a graph made, highlighted by when Civilizations were destroyed or capitulated

The above shows when each combat occurred in the game, counting each combat as a single instance.

This is the same graph as shown above, except instead of counting each combat as one, the combat is weighted by my units current Strength

The above graph indicates each combat (just like the other one), but weights each combat by my unit's current Strength rating.

Since I used Praetorians almost exclusively, there is not much of a difference between the two graphs. Perhaps in a later game (say when I go to Tanks), that will change.

Score

I really kicked the snot out of this game.

017-roman-war-machine-score.txt
is the dataset on which the following graphs are based.

Two lines sloping upwards indicating the time at which opponents were elimiated, this on in step wise function, at each even
As Events
The same as the previous, except over time, I do not see much of interest in these graphs, my score continued on upwards at a more or less linear rate
Over Time


Final Score:   2,726
 Normalized: 161,715
I recorded the Raw & Normalized Score on the Turn I defeated each opponent. That is on the the turn they either became my Vassal or were Destroyed by me (but not when others did the same).

{Spain was actually destroyed on 715AD, but I neglected to record the data before deleting all the old games. So, accurate data collection may not be my strong suit.}

The lines on both graphs start and stop at the same point, as that is a function of the graphing program I used. Or in other words, the lines are normalized to one another.

The slight bump at the very end is from a last ditch run for the finish line, which saw me capture four cities in the final turn, because... Why not?

These graphs basically tell me nothing. And as such, I doubt I will bother compiling them again.

Or at least, that is my first thought.

On second thought, they did tell me that my progress was slow but ever improving. I did not hit a plateau or fall (meaningfully) backwards, which is often the case. More cities and more land (from the War Machine) equates to an ever higher score. So, similar graphs may be of use in the future to identify points of failure and backwards decline. Still, I doubt I will undertake the effort, since the production of these graphs was done by hand.

Combat Statistics

When in doubt, look at the raw data.

I learned quite a bit (about my gaming style and what I need to improve), when I looked at the raw data that went into the Combat Graphs (from two sections above).


Some Basic Stats

Turns: 448
Combat Events: 286
Direct Conquers: 12 Civs

Most Combats in a Turn: 16
Most Strength in a Turn: 85

{In the above and the discussions that follow, Strength refers to a Summation of Unit Strength. So in the above (on the most violent turn), 16 separate Combats took place, involving a total (current) Unit Strength of 85.

Now, I've been know to go the long way around. And rest assured (in a moment), I will I. But almost everything I have to say can be extrapolated from the data above, after one realizes that a Praetorian fought in almost every Combat this game.
85 / 16 = 5.31
Meaning, in the biggest battle of this game, my Praetorians walked into battle with an average Strength of 5.31 out of 8.00... or at 66.4% of full Strength.

It gets worse.


Combat by Number
WinLose
Attack19243
Defence2229
     
Combats by Strength
WinLose
Attack104195
Defence7583
     
Percent Full
WinLose
Attack67.8%27.6%
Defence42.6%35.8%


The first table is a count of total Combat Events as split by Attack or Defence and my Winning or Losing. The second takes the same four way split, but looks at the total Combat Strength (of my units), instead. So a full health Praetorian would count as eight. The final table divides the number of events by the total Strength (in each quadrant) further dividing by eight to normalize the results for a Praetorian (almost all of these Combats involved Praetorians on my side and those that didn't were for War Elephants, which, also, have a Strength of eight). So, the last table shows at what percent of a Full Strength Praetorian I typically employed in each type of Combat.

Now, this would not matter (in the least) if the percents were all equal. But they are not. And a quick look at the charts will show that the bulk of my losses (43 vs 29) happen at my initiative (I am the one attacking), while utilizing weak troops (67.8% of Full Strength on the Wins vs 27.6% on the Losses). So, many of my Losses are from fighting with weak Units.

Oddly, when I dug down into the raw data (that's what the logfile is for), I found that the odds I was taking in these losing battles wasn't always horrible. The following being the first two losing Combat entries in the logfile:


While attacking in Celtic territory at Bibracte, Praetorian 5 (Crab Cake Cove) loses to Celtic Archer (0.24/3) (Prob Victory: 74.8%)

While attacking in German territory at Munich, Praetorian 9 (Iron Dam) loses to German Archer (1.74/3) (Prob Victory: 66.9%)


So, while not great odds, they are not horrible odds. And in reality, the only problem I see is that I lost, when I most likely would not have if my Praetorian had been at full strength... rather than at 0.24 and 1.74, respectively.

Once again, I have no option but to conclude:
  1. I am not maintaining Stack Discipline.
  2. I am not maintaining Health Discipline.
One cannot avoid all losses. But I am willing to assume half of my loses were the direct result of poor leadership (user error), which comes out to (90*(43+29)/2=3,240) 3,250 Wasted Hammers!

Such a number is intolerable... especially after one considers that every aspect of this game has led me to the conclusion that the computer simply does not know how to fight a war. And as such, I should be Playing with my Prey, taking my time destroying the AI, and milking it for all the Techs, Tribute, Wonders, Religions, and developed Cities that I can, which is exactly what I intend to do in my next game.

Let's Milk It!