1) Your overall score is the Sum of depreciated scores divided by the number of games to the 0.4 power. Your game scores depreciate 5% a month for 7 months (down to a minimum of 65% of their initial value). These are summed and divided by the number of game to the 0.4 power. Note that the 0.5 power is the square root and so dividing by about the square root of the number of games is a compromise between the average (which is sum divided by number of games) and the cumulative total (which is just the sum).

Basically the more games you play the less contribution each new game has towards your overall score. The following list shows what percentage your previous games contribute to your overall score and what percentage of credit you get for each additional new game based on the number of games submitted. This shows that once you've submitted 10 games that 96.3% of your total score is due to the scores of your first ten games and if you submit another game you will only increase your overall score by 38% of the 11th games score.

This says that each new game has less of an effect than the games that have gone before but that if you score 100K on your 51st game your overall total will still go up by 20K. So to answer your question the submission of any new game no matter how low the score cannot cause your overall score to go down. Note that your overall score will go down over time due to the depreciation factor of 5%.

The bottom line is that to eventually get the best overall score you want to avoid submitting low scoring games but they don't ever make your overall score go down.

n - Old Score Percentage (New Game Percentage)
1 - 75.8% (75.8%)
2 - 85.0% (64.4%)
3 - 89.1% (57.4%)
4 - 91.5% (52.5%)
5 - 93.0% (48.8%)
6 - 94.0% (45.9%)
7 - 94.8% (43.5%)
8 - 95.4% (41.5%)
9 - 95.9% (39.8%)
10 - 96.3% (38.3%)
11 - 96.6% (37.0%)
12 - 96.8% (35.8%)
13 - 97.1% (34.8%)
14 - 97.3% (33.9%)
15 - 97.5% (33.0%)
16 - 97.6% (32.2%)
17 - 97.7% (31.5%)
18 - 97.9% (30.8%)
19 - 98.0% (30.2%)
20 - 98.1% (29.6%)
21 - 98.2% (29.0%)
22 - 98.2% (28.5%)
23 - 98.3% (28.0%)
24 - 98.4% (27.6%)
25 - 98.4% (27.2%)
26 - 98.5% (26.8%)
27 - 98.6% (26.4%)
28 - 98.6% (26.0%)
29 - 98.7% (25.7%)
30 - 98.7% (25.3%)
31 - 98.7% (25.0%)
32 - 98.8% (24.7%)
33 - 98.8% (24.4%)
34 - 98.8% (24.1%)
35 - 98.9% (23.8%)
36 - 98.9% (23.6%)
37 - 98.9% (23.3%)
38 - 99.0% (23.1%)
39 - 99.0% (22.9%)
40 - 99.0% (22.6%)
41 - 99.0% (22.4%)
42 - 99.1% (22.2%)
43 - 99.1% (22.0%)
44 - 99.1% (21.8%)
45 - 99.1% (21.6%)
46 - 99.1% (21.4%)
47 - 99.2% (21.3%)
48 - 99.2% (21.1%)
49 - 99.2% (20.9%)
50 - 99.2% (20.7%)

2) Yes but not by all that much. I really haven't heard a specific percentage but I would be surprised if the difference between suicidal and cakewalk was as great as 2 to 1 and it's probably much smaller.

3) Yes. Exactly. Basically the intergral of the curve however there is an extra division by the turn number. That is that each turn the value of the curve is divided by the turn number squared. If it was the sum of values divided simply by the turn number this would essentially be the average (well not exactly but on the order of the average). In fact way back when the score was initially divided only by the turn number until someone submitted a 209 year game that scored 6 million points simply by hitting end turn forever. This caused substantial consternation and the recalculation of scores but that was before my time. If you search hard enough you can still find traces of this in the forums.

Note that the curves we're talking about are pop, income, military and research. There's some suspicion that influence may also come into play with pop to form the social score but that's pretty unclear and even if it does the pop term seems to dominate.

One other interesting point is that you can see the evidence of the extra division by a factor of turn number even today. When you get the score page it shows a score for each of the four factors; social, military, income and research and the total score. I believe the individual scores are "raw" scores that only have division by turn number and that only the cumulative total has the extra division by another factor of the turn number.

This can be used to gauge your performance. For better players that develop high values of each of the four components of score early in the game you will find that the overall score is greater than the sum of the four components of score whereas a lesser player that still may be able to generate a high score will have an overall score that is less than the sum of the four components of score.

The kinds of values I've seen are on the order of 2 to 1. So what I would call a "good" player would have an overall score that was twice as great as the sum of the four individual components of score whereas a "poor" player may have an overall score that is half that of the sum of the four individual components of score. This highlights that it's important to get high levels of income, pop, military and research as early as possible. I really can't stress the enough.

Basically just try it and see. End the game in December each year and see what you get for score. Do this a time or two and you get a feel for it.

You should do this yourself because it is instructive but the short answer is depends on what's going on in the game. As long as things are dramatically increasing then your score will be getting higher. By that I mean if your population is still growing, if your income is rising, if you continually building more ships then your score will continue to rise. However once your population tops out then so does your income and while you can generally continue to produce ships for quite some time before you begin to choke on the ships maintenance costs you quickly get to the point where you can no longer produce ships faster than you did the turn before. Once all this happens you quickly get to the point of diminishing returns.

Generally within one game year of when your population stops increasing is pretty much the end. This assumes that you've previously captured and maxed out all the resources of the galaxy. If you don't have all the economy, military and research resources in the galaxy then clearly there are good increases to be had by capturing them. However, in line with getting your points as early as possible the good player would have long since had all of the galactic resources maximized.

I've never seen the score actually drop but I have seen huge differences between the rate of growth from one year to the next.

Basically as I explained early points count more than late points. I've particularly noticed this in some of the Metverse League games. These games are in tiny to medium galaxies and so it takes much longer to build an array of military SB's and put a bunch of ships under it. This tends to quite significantly delay one of the major source of points in a game. Anyway in one particular game I scored 39.5K in a tiny galaxy game in 8 reported years, and I earned about 12K for that final year. I did look at how much I would get by extendeing the game until 9 years and the total was about 42K, clearly not worth the extra year, but as you can guess there was nothing growing in that last year, I had reached the limit of the ships maintenance that I could support, my pop was at it's max and my income was maxed as well. Nothing was growing and so another year gave me 2.5K when the previous year before my military was still growing and had contributed about 1/4th of my total score.

In my normal gigantic game I've never seen a benefit of going past 8 years although the final year can be a quite significant contribution to your final score. Like I said, try it and see, but once the graphs stop growing so does the score.

Seems to me as if your volunteering.

I have no real clue as to how the wiki works although I think anyone can contribute. I assume that Iztok Bitenc has probably made a fair amount of contribution to the wiki but that's a total guess on my part.

I also assume that most of the stuff there initally came from the developers, but again that's a total guess.

I think Purge once said on large or greater maps with planets common or greater, he passes (at least early in the game) on researching the extreme colonizing techs. Instead he favors letting the AIs colonize them and in the process over overextend themselves. He picks up the colonization techs when he invades the AIs.

Mumblefratz, during the colonization rush on a large or greater map with planets common or greater, do you bother researching the colonization techs?

Tech spending is one of the four components of score but as far as I've been able to determine it's really the weak sister of the four.

The quick answer is that if by taking money away from tech spending you can cause an equivilent increase in any of the other three components of score (population, income or military rating) then it is in your best interest to do so.

Basically I think this is directly related do to the scale of the values that can be achieved in each of the four components. I spent a number of games early on with a very strong focus on tech spending. I'm sure that I have achieved levels of tech spending that no one else has come close to and the resultant tech score was very disapointing. To give you an idea of the levels I went to I once created 6 separate PQ32 tech planets in a single game each surrounded by 16 economic starbase arrays and the level of tech spending that I achieved per turn was in excess of 200,000 RP's per week. I would have to guess that an ordinary gigantic abundant all type game using an all-factory strat might achieve tech spending levels of around 20K per week if that.

When I speak of the scale of the components of score note that my 200K RP per week research spending compares to my 1.6M BC per week income, my 10~12 million military rating (using 24 military SB's and 17~25K ships) and my 7~8T population. On the assumption that these are all "to scale" you can see how anemic 200K research is compared to 1.6M income or 10M military rating.

One important key to notice is the "wrap" that occurs in the timeline graphs. This can be seen in the timeline graphs for population and military rating under the civilization manager. Note that these same graphs can be displayed at the bottom center of the main screen and if viewed there, there is no display wrap. If you haven't noticed this wrap then perhaps you've not had high enough populations or military might.

The population "wrap" occurs at about 6.2 Trillion population. If anyone is familiar with an exponential growth curve (or an RC voltage curve, same thing) then what you see is the gradual rise in population that slows as it approaches it's asymptotic limit. As this curve reaches the 6.2T limit it instantly drops back down to "zero" and then continues on the same exponential path only offest by about 6.2T.

I have seen precisely this same behavior in the military rating timeline graph only instead of it occuring at 6.2 trillion it occurs around 6 million. Also if you're good and can achieve a military rating in excess of 12 million you can see that there is a second "wrap" that occurs at 12 million as well. Note that I've never seen a comparable wrap on either the income or the research curves.

It's my supposition that a population of 6 trillion would generate a comparable score to an income of 6 million per week or a military rating of 6 million or a reseach spending of 6 million. This is what I mean about scale. Again all this is assumption on my part but if the best possible performance you can get on research is 200K per week how valuable can that be if this is ranked on the same scale as 1.6M bc's per week or a 10M military rating.

The component breakdowns of my scores seem to bear out these assumptions. I can get 1 wrap performance on population, double wrap performance on military rating and no wrap on either income or research. Also it seems that all components are scored logarithmically, that is to double the score you need at least four times the value or in other words linear increases in score require exponential increases in value.

That is the rub. I believe that you can never get rid of a lot of the micromanagement and that you actually really don't want to get rid of it all (otherwise you'd be playing MOO3).

However, there are many ways to reduce the micromanagement aspect of a lot of these things. They're hard to descibe but wherever you can find ways to streamline and "assembly-line" the process the better you are.

In fact the way I build lots of ships is far more micromanagement than what you go through. From your explanation it's clear that you allow your planets to directly build the ships for your military SB array. While that's pretty much required because of economies of scale in our MVL games for example that's not how I do it in my gigantic games. In those games I maximize both my income and the number of ships that I can produce as well as how many I can afford to maintain by converting *all* of my planets to 100% stock markets and buying ships directly. This sounds like it's very inefficient and from the point of view of the cost of producing them versus the cost of buying them it is, but from a scoring perspective it's a win-win scenario. I get a much higher income and I produce ships faster.

You get no points for industry but you do get points for income so maximizing my income is a means to an end all by itself. Then in conjunction with the 6 economic resources that a gigantic galaxy usually provides you end up with 5X (assuming all tech and the MCC as well) the income you otherwise would have. If you're lucky enough to get the economic prosperity event then this doubles to 10X. This pretty much automatically covers the increase in cost of buy versus produce and since you have all these stock markets instead of industry you can pretty much produce the same amount of ships but with a much higher income and therefore a higher score.

One key is the "assembly line" of the tedium. For example buying ships from the colony list in your civilization manager can be done in two ways. One directly from "buy ship" within the civilization manager itself. The other is to go from the colony list to the planet screen to the starport screen. Even though this takes more clicks it's actually faster. However there's an even faster way and that is to sequence through your planets using the X or Z keystroke commands from the main screen (or just go through sector by sector) and directly buy the ship from the main display. Once you have a lot of planets "producing" ships this turns out to be much faster. Then I use a colony rally point for all my planets just to get the ships auto launched but each turn I then use my governor to turn off the autopilot destination. I then wait 7 or 8 turns dependent on my logistics and fleet up all the ships at each planet and send them off to my rally point only once every 7 or 8 turns. This actually saves a lot of time and it also saves a lot of memory over having ships flying off independently.

There are many ways to do things like this. I've seen people use a mixed build/buy ship strategy where they will produce the hull from industry and upgrade the ship instead of totally buying it. This is a lot more cost efficient and it can also be "assembly-lined" by alternating the ship built each turn using the governors and then allowing one turns ships to autopilot to the rally point while you're doing the upgrade on the other class ship. Of course both ships are identical only having separate names so that they can be upgraded idependently. I've also seen this combined where you do a whole bunch of upgrades that put you in debt for a turn or two so that you only have to deal with it every other or every third turn.

If you give the entire process some thought then you can come up with ways to make the tedium easier and this is another method that you can differentiate your games from everyone else's. There really are no secrets to score, everyone knows pretty much everything that can be done, the question is how much tedium can you stand and secondly how much tedium can you eliminate by careful consideration of the processes you go through.

Thanks for your kind words but it always seems to me that the advice I give is rather hazy and indefinite. There are just so many details that really make any particular strategy work that are just impossible to convey.

What I try to do is give a big picture and hopefully some parts of anything I say will make some sense and fit in with at least some part of everyone's own style of play. I think that's the best that anyone can hope for. I'm still trying to digest and incorporate things that Magnumaniac told me long ago. I've never been able to duplicate his colony rush but then I don't think that anyone else has either.

Wait. I thought this was a game! What about fun!



He likely used the Yor super ability to slow the AI colony ships so that he reached the planets first. Still, most impressive!

just a thought about Mag's colony rush, and commenting on Mascrinthus...

true he could use the Yor SA in DA to his advantage but he perfected that strategy in DL first before moving onto DA, so I doubt it had that big of an effect.

Im wondering if he planet hopped. Smart researching with the Yor's miniaturization bonuses could produce very fast and cheap colony ships right from the start. Take your colony ship and send it to Planet B. Colonize. Buy/build starport (anywhere from 1-5wks) and then buy/build the cheap colonizer. Send off, even with non-optimal population, to Planet C. Repeat.

Off the top of my head its the only method that I could see being fast enough. Couple this with a good population and economic bonus and you wouldn't even to worry about going bankrupt. Also, all the worlds would add to your research points automatically so that wouldn't be an issue either (falling behind I mean)

well...as I let my mind wander here....

like i said, a good population growth rate and economic bonus would go a long way. Or being super breeder in DA.

and then, really knowing your tech lines. I think this is a must for being able to compete with the AI, esp since they (the AI) tends to be more scattered in its research; they can't focus on a tech line like a human player can.

esp with the manufacturing buildings (unless you have the Hive ability) using the lower end structures (basic factory and factory) to get a start on your worlds; Manu Centers and Industrial Sectors take way too long to build from scratch.

In my mind, I could picture a way of doing it, with actually productive worlds. But you would really need to do some micromanagement and be conscious of everything that was going on. And really thats a level of play that the average player is either not willing or not capable of achieving or maintaining.

I do remember him discussing using the Yor's super ability in DA to drop a line of planets to stop the other AI's colony rush dead in it's track and it does make sense that you could do this. I was more referring to his normal run of the mill performance in DL where the differentiation between races was not so severe as in DA.

But I do understand and have been using the idea of "farming" AI's by intentionally letting them develop planets for me before taking them over. I've found this to be a particularly useful strategy in the tiny to medium galaxy games within the MVL. A lot of the gigantiac game things don't really scale to these small games but immediately jumping on opponents that represent a legitimate threat while intentionally letting weaker ones develop is something that lets you get a leg up in these smaller games. Again it's not as dramatic as the super ability "farming" that you could conceivably do in DA, just a matter of jumping on the Torians and Drengin while giving the Korx and Iconians free rein.