The truth about Military Rating

Published on August 5, 2008 By Mumblefratz In Metaverse

It's surprising how long one can play this game without gaining a true understanding of how certain things work.

Without going into any great detail I probably have about as much experience as anyone in maximizing the military component of my score and you'd think that by now I would have a good working understanding of how it works. However, I've just recently discovered that this assumption is totally wrong.

I've always worked on the basis that the military component of your score (like all other components of score) is merely the value of your military rating divided by the turn number squared summed each turn over the course of the game. There also may or may not be some "constant of proportionality" that linearly scales this value.

This is the way I believe military scoring works, which is the same as how the other components of score (income, population and tech spending) work. Basically each of the four components of score are the "front-end weighted", "area under the curve" as displayed in the population, economy, military and research timeline graphs under your civilization manager.

I've written about this many times before and to my knowledge this is consistent with how all the top players view the scoring system. But while I still believe that this basic framework is true there is an implicit assumption here that has been unquestioned for a very long time and in fact this unspoken assumption is *not* true.

This unspoken assumption is about how your military rating is calculated. Looking at the other three components of score it's very obvious what determines your income and population graphs. It's obvious that the value of each of these two graphs at any one point is simply your gross income and your total population. There is no scaling or other "calculation" required; the relationship is direct and one to one. In the case of your tech score it's only slightly more complicated. It's actually not the total income spent on research but in this case I believe that the value of the research graph at any one point is simply the total amount of RP's that you produced that turn which takes into account all possible sources of research bonus.

The natural assumption is that your military rating is treated as above, that the value of your military curve at any point is simply the value of your military rating at that turn. I still believe this to be true as well. However what is definitely *not* true is how we've been told our military rating is determined.

To the best of my knowledge I've always been told that your military rating is the sum of the attack value plus defense value plus hit points divided by 10 of all of your ships once all source of bonus are included. In other words:

Military Rating = the sum of Attack + Defense + HP/10 of all ships

While I do believe the above to be partially true this gives results that are wrong by orders of magnitude. Basically I've come to the conclusion that there has to be some function of the number of ships in the denominator of the above equation.

It's not so surprising that this assumption could go unquestioned for so long. The number and value of ships that people have at any one time is constantly changing. So are your weapons and defense bonuses which are usually different to further confuse the issue. Add to this having some ships in orbit around the Spin Control Center and just having some ships in orbit for the 25% bonus. With all these effects taken into consideration it's not surprising that no one has really taken the effort and tried to correlate some abstract calculation with the displayed value of your military rating.

Like most everything else in this game it's generally not necessary to know precisely how some equation works, it's usually enough to know that more is better. However when trying to maximize score sometimes it's important to know how much more. While more is always better it's generally not true that twice as much is twice as good. It becomes important to know these things so that you can determine where to best put your effort.

Anyway I don't yet have an answer to how your military rating is really calculated. I do think that I have enough data in the many different save games I have to make some kind of determination but I also believe that it's not necessarily an obvious relationship and so I would appreciate any possible help in understanding this.

I don't currently have access to all my data but I'll end this OP with a couple of examples that prove my claim that military rating is somehow inversely proportional to the total number of ships that you have.

My first example can be demonstrated with a screenshot that I'll post later but basically the method that I use to build and move ships to under my military SB array goes as follows. Each turn I buy ships at a number of planets which have been set to autolaunch to a rally point, but each turn I go into the governor and kill the rally point destination. I then wait 7 or 8 turns letting the ships sit next to the planet that built them until I have a full fleet of ships at which point I fleet them up and send them off to under my military SB array. This dramatically reduces the tedium of building and moving ships.

The point is that each turn as I build another 300 or so 1/1 fighters my military rating actually noticeably goes *down*. This proves that your military rating is somehow inversely proportional to the number of ships because as these 1/1 fighters are built they don't increase your attack/defense values very much (i.e. the numerator of the equation) but do increase the number of ships (i.e. the denominator). Then once these new ships are moved to under your SB array then your rating shoots up again because the military SB array dramatically increases the value of these ships offsetting the effect of the ship number increase in the denominator. The point is no matter how small an increase in attack/defense a ship is, if your military rating is not an inverse function of the number of ships then your military rating can *never* go down. It may not increase very much or may appear to stay the same on some larger scale, but it could never actually decrease as you build more ships.

The other example is the one that actually caused me to question this in the first place and can make a difference in how you might decide to play the game. This came about in my quest to become the first to achieve a 1 million point DL game.

My normal course of action is to conquer the galaxy, build up my population and income, build a military SB array and then build about 17K 1/1 huge hull fighters for under the array and then end the game by upgrading all of my 1/1 huge hull fighters to 25/450 BHE/ZPA dreadnoughts. The reason for 17K ships is that on my machine that appears to be the limit to the number of ships that I can upgrade at one time without running out of memory. It's also a convenient point in that by the time I've built that many ships my population and income have pretty much reached their maximum and beyond that point system slowdowns become intolerable.

My problem has been that this has gotten me close to a 1 million point game but I've never quite reached it. In my latest game it appeared that this strategy would result in a game somewhere in the 850K-900K range. Again close, but no cigar. So I tried another approach. In the 2nd case I instead built about 32K ships, 17K that I intended to upgrade per usual and another 15K or so just to increase my military score above the "hump" (or so I thought).

However an interesting thing, or very annoying thing dependent on your point of view, happened. In the first case when I just built 17K ships my military rating was about 5.7 million prior to the upgrade and then 14.2 million after the upgrade. So an upgrade of 17K ships gave me an increased military rating of 8.5 million. In the second case my first indication was that with 32K ships my rating was only 9.2 million instead of the 10.7 million I would expect if military rating were linear. Secondly after the upgrade (in this case I could only upgrade 15K ships) my final military rating was only 15.7 million so the upgrade of 15K ships increased my military rating by 6.5 million. Again if military rating were linear then the upgrade of 15K ships should result in a military rating increase of 7.5 million instead of the 6.5 million actually achieved.

The whole point of this is that I spent a whole game year of building an extra 15K ships for a net increase of 1.5 million to my military rating and probably less than 50K increase to my score.

The point about really understanding how your military rating works is that this could result in a potentially better scoring strategy such as building far fewer ships (but perhaps earlier in the game) and then spending effort maximizing the score of some other component (like tech for example).

Anyway if anyone has any insight or experiences to offer related to the non linearity of military rating please contribute them. Also if I find out more detail from a number of data points, I'll try and figure out the true military rating formula and post it, because although what we've been told is probably true, it's certainly not complete.

Comments (Page 1)

VegJed

on Aug 05, 2008

Actually, I noticed fairly early on that there was a "diminishing returns" effect to the military rating. Granted, I've never played on the scale where building more ships could cause the score to go down, but around when I started using the SCC to boost my military score, I noticed that each "jump" in my military score was slightly smaller than the one before it. I saw it as slightly annoying, as my opponents military would be steadily increasing, so my periodic jumps would eventually not be enough to keep me in the lead, but other than that I didn't think much about it, at least not enough to comment on that fact on the forums.

It's just one of those things you have to take into account when you're playing. Playing on small scale games, as I pretty much exclusively do, it's more annoying that the military rating of a ship is only counted if it has an attack rating, so when I choose to go good in a game (as I often do in diplo games) I can't just fill the ships up with my super strong "good" defenses, I have to throw on at least one laser.

BarneyStinson

on Aug 05, 2008

I've always felt in the back of my mind that the number of ships I had was a determining factor in my military score, but I never had anything to back it up beyond my "hunch".

What you're saying makes sense. Perhaps the reason #/ships is factored in is to lessen the gains from strategies such as the 1/1 starbase-magnified smaller ships.

As I mostly play small or medium galaxies, my idea of a good military rating is around 30,000, nowhere near millions, so there isn't as substantial a disparity in the math for me to truly notice something else at work.

Mumblefratz

on Aug 05, 2008

What you're saying makes sense. Perhaps the reason #/ships is factored in is to lessen the gains from strategies such as the 1/1 starbase-magnified smaller ships.

Probably correct and to a large extent I view this as a positive rather than a negative. Truth be told I'm sure most folks that go for score would prefer not to have to go the 1/1 fighter route but generally feel compelled to do so because otherwise you're at a disadvantage.

I've always viewed a *reasonable* amount of this as OK but really do prefer to cut it off at the first sign of diminishing returns. I know numbers like 17K ships sound like a ridiculous amount but at the incomes and number of ships per turn achievable in a gigantic game this really only represents a bit less than 1 year (48 turns) of ship building.

Another point that seems improved in TA is the necessity of building a military SB array. This may in fact become obsolete in TA because higher levels of weapons and defense bonus are achievable in the immense galaxy. Once ships reach levels where the 25% bonus you get from ships being in orbit is close to the levels you get from a SB array then you may as well skip the array and just build high quality ships to leave in orbit. I think most people that feel compelled to build arrays would prefer this.

The bottom line of finding out that your military rating is inversely proportional to the number of ships that you have is that it tends to put more focus onto quality rather than quantity and I believe that to be a good thing.

BarneyStinson

on Aug 06, 2008

Probably correct and to a large extent I view this as a positive rather than a negative.

Yeah I agree with you there.

Something I'm curious about is this: I see people with scores in the several hundred thousands per game. Are these games indicative of using these score-boosting methods, or because of a massive, planet filled galaxy? (or both)

Let's say TA game, gigantic galaxy, all abundant. 8 year completion, no artificial score inflation, conquest victory. What would be the estimated score, ~350,000? How much extra score would the military rating-boosting methods give? I'm trying to crack the top 25 and players with one monster scoring game of 500,000 or 600,000 annihilate my dozens of games of 35,000-45,000. Am I doomed to playing tons of more games, or am I forced to resort to these methods to boost my ranking to that level? I don't like doing that stuff. I like just kicking major A in a game, doing the best I can to obliterate the enemy, and getting a great score that way. Is there hope?

KzintiPatriarch

on Aug 06, 2008

I've always felt in the back of my mind that the number of ships I had was a determining factor in my military score

I concur...it's something I've suspected for a while. I discovered it not out of curiosity, but necessity. My computer cant handle nearly 17k ships, so I build maxed-out huge hulls for the array right from the beginning. My most recent score is 508k on a huge all-abundant map, which is at the limit of what my system can currently process.

Sentient species taste better...

Mumblefratz

on Aug 06, 2008

Something I'm curious about is this: I see people with scores in the several hundred thousands per game. Are these games indicative of using these score-boosting methods, or because of a massive, planet filled galaxy? (or both)

Basically to hit the max score you need to do *everything* and of course that would include any and all "score boosting methods" known and imagined.

Clearly you want as many planets and mining resources as possible. You also want the fastest tech. Generally most people just use a shorthand and call these "abundant all" type games because that's literally what you want, everything you can get.

Ultimately what very well determines how high any particular game can score is the number and type of mining resources there are in the galaxy. In DL the average is 6 per type for a total of 30 but the most important are economic and military followed by research. Morale are nice too but usually having more of them is just a convenience whereas as far as I can tell influence are useless for scoring purposes.

As far as "score boosting methods" one of the most effective is to build up your pop, income and military as soon as possible. For the top scores this usually means that the galaxy has essentially been completely conquered before the end of the second year of the game. From there it's more a matter of keeping your approval high so that your pop grows fast and getting all the resources of the galaxy fully maxed out. Your income follows from your pop and your ability to build up huge quantities of ships follows from your income.

Let's say TA game, gigantic galaxy, all abundant. 8 year completion, no artificial score inflation, conquest victory. What would be the estimated score, ~350,000? How much extra score would the military rating-boosting methods give?

No real clue here. I would have to take "no artificial score inflation" to mean ending the game at the first instant that you possibly could without even allowing your pop and income to reach their final values let alone do any military buildup. In this case it really depends. If you can conquer the galaxy within a single year you could probably still achieve 400~500K. If it takes you 3~4 years to conquer the galaxy you might be talking about 200~300K. If it really takes you 8 game years just to conquer the galaxy then your score would probably be in the 100~200K range.

The same thing with the military buildup. If completed early enough it could add as much as 300~400K by itself, if completed 4~5 years into the game it may only add another 100~200K. The thing is you cannot over-estimate the importance of getting things done early in the game. I've seen cases where 5K ships built early can outscore 50K ships built late. It's not just a matter of what you do, it's when you do it that will have a bigger impact. That's why you can do exactly the same things I do and may only score 300K while I'll score 900K.

Basically people look on these "score boosting methods" as a panacea that will allow anyone to score any particular value they want ... they're not. Doing them will always improve your score over not doing them but usually not by as much as you would hope and most likely there will still be people that can not use them and score higher than some that do use them. However it is clear that to excel at the highest levels these are basically a requirement. You can't achieve the highest scoring games by leaving points on the floor.

Mumblefratz

on Aug 06, 2008

My most recent score is 508k on a huge all-abundant map, which is at the limit of what my system can currently process.

Seems like a pretty good score for a huge galaxy, it's the highest huge that I've heard of, although to be honest I haven't really gone looking much.

Certainly scaling the size of the game you play to the capability of your system is a reasonable thing to do but I think the real key is memory size. You definitely need 2GB of RAM to not crash out at the 17K upgrade level. However, this new understanding of value per ship as an important requirement lessens the value of quantity over quality at all costs. As your game points out fewer high quality ships earlier may in fact be just as effective as more high quality ships later in the game.

I know that Mag essentially used the same technique in his 1M TA game. He built only 5K top line ships but did so earlier in the game and by not doing a final upgrade was able to continue a huge amount of tech spending to much later in the game. While this resulted in a lower military score than could be achieved it resulted in a much higher tech score for an overall score in excess of 1M. The point is that there are different ways to get to the same result.

Piznit

on Aug 06, 2008

I wonder if the game equation for calculating military score is complex enough to see the growth rate of your military might as a factor.

Perhaps your score goes down for that 7 or 8 turn period while you get enough ships to fleet up because your growth rate decreases. Then, obviously, it jumps up again as all those ships fly under the array. What would happen to your military score if you built say, 10 ships increasing over time, then a few turns decreasing over time.

Example:

Turn 1: build/buy 1 ship.
Turn 2: build/buy 2 ships.
Turn 3: build/buy 3 ships.
Turn 4: build/buy 4 ships.
Turn 5: build/buy 3 ships.
Turn 6: build/buy 2 ships.
Turn 7: build/buy 1 ship.

Should be easy enough to see if score scales linearly or exponentially, as well as to see if score diminishes when you start building less ships.

Philocthetes

on Aug 06, 2008

I have a possibly stupid digression-question: Could the "truth" about the math here also be related to how the computer players are so bad at estimating the real effectiveness of the human player's military?

I'm still hung up on non-military wins (for a variety of reasons), and from DL through DA and TA I've consistently seen computer players very wrongly decide I was easy meat, and it appeared to me to be simply because they had far more hulls built than I had.

Those swarms of hulls are no good to them when only a handful of their latest and greatest designs can even scratch my defenses. But they call me unfit to survive and make me take the time to destroy all those wimpy hulls.

Mumblefratz

on Aug 06, 2008

Should be easy enough to see if score scales linearly or exponentially, as well as to see if score diminishes when you start building less ships.

I'm going to try a few experiments once I finish my current game, but I believe the dependency is small enough that you're going to need to be dealing with a lot more than 10's of ships. I have enough save games at various points to be able to get a number of widely separated data points that should allow me to hopefully get a good "fit".

However based on a very rough calculation the attack + defense + HP/value of 17K ships under a 24 SB array is about 26M and the corresponding military rating that I got at that single point was about 5.7M which implies a denominator of about 4.5 based on about 17K ships. Interestingly enough the (base 10) log of 17K is about 4.2 which is in the ballpark but if I have to make a guess without any real data I suspect the numerator is a fractional power of the number of ships. If so based on the single data point above it would imply the denominator is close to the number of ships to the 0.15 power.

One of the best tests that I can probably easily do is to find a save game with about 2000 ships not under the array and somewhere in the vicinity of 5 to 10K ships under the array and then delete the 2000 ships not under the array and see the precise values of military rating increase. Actually it would be nice if I can see this effect with unarmed ships as well and that way I can essentially graph how the military rating varies with the number of ships while essentially holding the total attack/defense/HP value constant.

However it probably won't be that important to know the precise formula. After all what practical value is there in knowing whether the denominator is the base 10 log of the number of ships or the number of ships to the 0.15 power? The point at which worthwhile to stop will still probably need to be empirically determined. There's still an unknown in all of this even if we know the exact formula and that unknown is the length of time needed to reach any particular quantity of ships.

We've all know that values early are worth more than values late so in addition to the knowledge of that there are diminishing returns in your military rating to increasing numbers of ships there are also the additional diminishing returns to your score that occur due to the time delay of building those additional ships.

To me the end result is to do as I've pretty much always done which is to take what the game gives you easily and then call it quits. There's no quicker way to increase your military score than by performing a "final" upgrade of as many ships as your system can support. The result of such an upgrade depends on a lot of things like your level of miniaturization, how many military mining resources you have, any extra weapons or defense racial abilities, etc. But to be able to triple your military rating from about 5 million to close to 15 million in a single turn "has" to be a cost efficient thing to do. But what I've found is that trying to go above that point is probably a waste of time and effort.

One final caveat emptor about all this scoring stuff. If you find it fun to do then fine, then it's simply a part of the game that you enjoy. But if it's not then there's no point in really forcing the issue because in the end if you don't find it fun you simply won't continue the game. In the end it's best to play the way you want to instead of how someone else happens to play. To each his own.

Mumblefratz

on Aug 06, 2008

Well the dependency on numbers of ships is so slight as to not really make much difference in this. This is still a result of the fact that having 1,000 ships may reduce the overall value of the sum of your attack/defense/HP by a factor of 2 over having only 10 ships but the increase in the numerator by going from 10 to 1,000 ships is a factor of 100. Clearly this is still very strongly favors quantity over quality.

However this does suggest an easy way to address the problem that AI has with quantity over quality as well as reduce the benefit of spamming huge numbers of ships and that is to *increase* this dependancy. Again this is a touchy-feely thing that would need to be "tweaked" and play tested but I suspect that increasing the denominator to something like the square root of the number of ships would be a benefit to both of these things. In this case if your opponent has 100 times the number of your ships but your ships were 10 times more powerful than his then the square root denominator would call that "even". That's probably closer to reality than the current system.

In all honesty as an acknowledged score monger I build military SB arrays and put a bunch of ships under them more because it's what I *have* to do to compete as opposed to doing it because it's something I *prefer* to do.

In some sense military SB arrays are to GalCiv2 what steriods/HGH is/was to baseball. You need them to simply compete. However if suddenly they were no longer effective I would doubt that many would miss them and that most folks that use them would actually be relieved.

I've made the statement before that no scoring system is perfect but that ideally the scoring system should properly value activities that are indicitive of superior play and therefore more closely correlate score to skill. Clearly the ability to produce quantities of high quality ships quickly *is* indicitive of superior play but to be able to extend this forever is *not*.

Livonya

on Aug 06, 2008

Mumblefratz -

I definitely think that there is diminishing return for ships.

This is one thing I learned from my million point game where I was forced to build 50K ships due to the fact that I could not do more research spending.

I am working on a new game that I hope to have done in 2 weeks.

I am doing something a bit different with this game.

I really think the key to high scores is very quick victory, and a balanced approach between military and tech spending.

We will see. I should clear 1 million again, but I am hoping to get closer to 1.5 million... keeping my fingers crossed.

- Livonya

Mumblefratz

on Aug 06, 2008

I really think the key to high scores is very quick victory, and a balanced approach between military and tech spending.

That's certainly what Mag did in his 1M point TA game. Although by quick victory I assume you mean conquering the galaxy very quickly but still taking the time to max out pop, income, etc. After all, even Mag's 1M game took 7 reported years even though I would be surprised if it took him much longer than the end of year 0 to actually conquer the lions share of the galaxy.

Actually it was your game with the 50K ships and a 10M military rating versus Mag's game with 5K ships and a 5M military rating that got me started thinking down this path. My case is somewhere between these two extremes with 17K 1/1 fighters worth a 5M military rating prior to the upgrade and worth a 13M military rating once upgraded.

I still think that I got a bit more out of 17K ships plus upgrade than Mag got out of 5K ships plus his 400K per turn research spending. While it's hard to compare DL versus TA games because of so many basic differences, I at least can see how each approach could be reasonably argued.

Certainly for DL with it's significantly fewer planets and hence lower populations, I don't believe the balanced approach will ever get you close to the 1M point goal. That's not to say that a balanced approach isn't more "efficient" even in DL it's just saying that balanced won't get you to 1M in DL.

Still like I said before, it's hard to argue that the almost tripling of your military rating in a single turn that you can get from such an upgrade isn't "efficient" even if it does spell the end of all research spending. But I do think this implies that going beyond the point of the largest upgrade you can manage *is* definitely bucking the efficiency trend.

Livonya

on Aug 06, 2008

Yeah, by quick victory I meant owning 99% of the planets as quickly as possible.

The actual game should not end until year 7 or 8 or whenever gains no longer are coming from increased time.

I have yet to try your method of bankrupting your society by updating 17K ships in 1 turn. That is still on my agenda. I think I will try that next. It sounds like a very efficient use of real time, as the rest of the game is simply pressing end turn. Very nice use of real time.

I have never seen a post where Mag said he was spending 400K per turn on research spending. Is that an exact figure or just a guesstimate by you? Do you have a link for that?

The most I have ever spent on research spending is roughly 60K per turn, and I ended up getting about 200K for that in the final score.

I am about to stop building in my current game and switch to 100% research spending and quick buying of ships. I am not sure how much my total research spending will be, but I do know my main research planet will be turning out 20K research points per turn.

- Livonya

Mumblefratz

on Aug 06, 2008

I have never seen a post where Mag said he was spending 400K per turn on research spending. Is that an exact figure or just a guesstimate by you? Do you have a link for that?

It's a direct quote but it was most likely made at the Tyranny of Evil site in the private empire forum.

As I think I've mentioned we'd love to have you join us and I'm sure there's all sorts of useful information that we'd be happy to share, most of which you probably know but there might be a trick or two we have up our sleeve.

Anyway, I'm pretty sure his basic strategy was to buy the 5K top line ships on time as early in the game as possible which used up about half of his economy in lease payments and ship support and then spend the rest of his income on research spending. Part of this strategy took advantage of some of the idiosyncracies in the different tech trees for different races in TA and being able to trade for certain other races technologies. I still think this applies to DA as well although some benefits might not be as significant there.

I tried the research route in DL and the benefits were not overly impressive. In my case I primarily went with 6 PQ32 uber research planets each boosted by 16 econ SB's but I think I topped out with a total research spending of 200K and far less benefit to total score than that. I think in Mag's case he simply overbuilt all industry with Discovery Spheres on all planets and had a more normal distribution of research although I'm pretty sure he ended up with some major resource mining research bonuses to help this out as well.

It is attractive to use income as a double duty source of score. In the first case your income earned counts directly in your economy score and then when spent on research it counts towards your tech score. However you do have to decide to forgo the upgrade path. The idea is that if you can get a 5M military rating for 5K top line ships and because you buy them on a lease and can probably do so far earlier in the game then you could ever hope to build 17K ships, the net effect to score is increased and then you can end out the game spending money on tech instead of ships. It's an interesting option and whether or not it turns out to be marginaly better or marginaly worse than the upgrade it's probably at least in the ballpark. Plus it just opens up the option of a different way to play for those that prefer to avoid the excessive ship route. At the very least it's a reasonable alternative.

Welcome Guest! Please take the time to register with us.
There are many great features available to you once you register, including:

Richer content, access to many features that are disabled for guests like commenting on the forums.
Access to a great community, with a massive database of many, many areas of interest.
Access to contests & subscription offers like exclusive emails.
It's simple, and FREE!