Social Effects Comparative Analysis
Contents |
Social effect weight idea
Social effects were not created equal. They were just created, you know. Obviously, some of them bound to be better than others. That is intuitively perceived by veteran players as well as game designers. Indeed, some society model effect mix can be very unequal. Take a “Free Market” for an example. This is due to difference in perceived weight of contributing social effects.
The purpose of this research is to calculate social effect weights to help out social engineering mod designers to compare and balance society models. The big idea under this calculation is to tie resulting social effect weight to basic measurable game variables such as number of bases, average base size, average tile yield, energy allocation, etc. Even though assumptions about these variables are still subjective they can be observed and statistics can be gathered easily. Once that base is established math will speak for itself. Of course, due to the nature of initial assumptions this still won’t give us an exact answer but it will confine our results in the more or less narrow range usable for society models evaluation.
Let me reiterate again that greater value of society model doesn't mean it should be used all the time. Game situation varies and so is society model tactical advantage. Yet when we see that one model on average is ten times less valuable comparing to others we can clearly conclude it is badly designed and will be rarely used if at all. That gives us clue which society models are subjects to change for improved variability.
Detailed explanation
Imagine you are choosing between two society models: A and B. They are same with only difference that A has +1 INDUSTRY, whereas B has +1 SUPPORT. Which one is better? Of course, the answer depends on current game situation as different SESocial Engineering tweak different game parameters. In the above case the +1 INDUSTRY essentially gives you +11% of minerals production. Whereas +1 SUPPORT saves you one mineral on support providing you have supported units. So mathematically, they deliver same benefit when you have 9 minerals base production on average. Below that SUPPORT is better and above that INDUSTRY is better. Simple. Now what if you are at +2 on SUPPORT scale already, your base size is larger than 4 and you have more than base size units supported? Then, clearly, +1 SUPPORT gets you to much better place in term of minerals saving. From the other hand, if you are already at +4 INDUSTRY instead, then additional step on INDUSTRY scale is worth twice as much. It would be interesting to know the ratio of +1 SUPPORT to +1 INDUSTRY averaged across possible games cases and stages. Our goal is not to find an exact statistical average but the perceived value of one SE over other in term of game play as it understood by human player. Like being in a war or not is two game cases human can clearly distinct. So we would need to evaluate how these two cases affect the above SE ratio. We need both variation range as well as middle of the variation interval.
One can then estimate Society Models value using the above calculated SE weights. SM value will be simple sum of all contributing effect changes with their respective weights. For example, vanilla Police State (+2 Support, +2 Police, -2 Efficiency) value would equate to +2 * <Support weight> +2 * <Police weight> - 2 * <Efficiency weight>. Then we can compare SM values across game stages and cases as well as get the global average. Taking in account that our calculated SE weight show how much change in this effect contributes to player advantage we can use calculated SM values as a measure of their attractiveness and, therefore, the prediction of how often they will be used. Social engineering mod designers then can use this scale to balance their creations.
Let me reiterate it once again that those values are not even meant to be exact. They give you a range at best. This is not an exercise in exact math but rather practical tool for SM modders. That's it. It is completely objective yet parameters selection is arbitrary and approximate. Feel free to add your own statics, assumptions, observations and/or correct mine. I will appreciate any constructive input.
Comparison technique
Social effects conversion
Thanks to complex game design almost any action of effect can be achieved by different means. That allows us to emulate one effect with another by varying some game parameters thus calculating their comparative value on the way. Indeed, if one effect gives you some benefits and another one (with some game parameters fiddling) gives you same then, clearly, they provide the same value. The relative social effect weight can be found then as a proportion between the benefit and the level of social effect increase. Of course, not all effect can be converted to each other that easy but we’ll figure it out on a way.
Averaging
Some social effects are greatly affected by certain game conditions. It would make sense to estimate social effect value under specific circumstances and then average across them. This would show us average effect weight as well as its variation range.
Keep in mind that assumption about averaging dimensions and values are completely subjective. I feel like small variations in these values will not drastically distort final results. In other words, I believe no social effect weight will suddenly jump out of charts just because of slight change in one game variable.
Game stage
One dimension is game progression stage: early, middle, late. Each stage more or less naturally defines other game variables such as number of bases, average base size, average tile yield, average income multipliers, etc. These variables grow smoothly through the game so considering just three points on a history line should be enough for good averaging. Below are stage description and their corresponding assumption about game variables I used in my calculations.
| parameter | early game | middle game | late game |
|---|---|---|---|
| turn range | 1-100 | 100-200 | 200-300 |
| number of bases | 6 | 12 | 24 |
| average distance from HQ | 3 | 5 | 7 |
| average base size | 3 | 5 | 7 |
| average base yield | 8-6-6 | 15-12-18 | 24-24-40 |
| Efficiency rating | 0 | 1 | 2 |
| Energy allocation | 3-0-7 | 3-1-6 | 3-2-5 |
| Energy multiplier | 1.0 | 1.5 | 2.0 |
| Mineral multiplier | 1.0 | 1.0 | 1.5 |
This is just a sample. See complete list in attached worksheet.
War and pacifism
This dimension is about war situation and pacifism drones. Even though war and pacifism do not always happen together they quite often do. These conditions can be on and off by player(s) desire. Some social effects behave differently under these conditions and need to be evaluated in both cases. Most notable social effects falling in this category are: Morale, Police (-3 and below), Planet (fighting other faction psi units), Probe (hostile actions).
Terms and helper variables
<energy> = average base raw energy yield. <economy> = energy reserves. This is to less confuse it with raw energy yield. <labs> = research reserves.
Since we are going to express everything with minerals we would encounter following “energy production to minerals production ratio” coefficient very often in calculations. I’ll use this coefficient to shorten formulas. Essentially it means how many minerals you can produce by directing all energy to reserves without any loss comparing to you actual mineral production.
E/M = (<energy> * <energy multiplier> / 2) / (<minerals> * <minerals multiplier>)
Another helper variable is level of Inefficiency = Max Efficiency – Efficiency. This variable is used to calculate energy loss due to unequal allocation, corruption, and b-drones.
Assumptions
Let me reiterate once again with big letter: I do not claim these assumptions to be right. This is just a starting point to run my calculations against. I would be more than happy to see how things change with different assumptions. In fact it'll be interesting to research how strong SE weights depends on certain input parameters. If they do not vary much it gives us assurance that calculated weights are indeed close to perceived values.
World size is normal.
Difficulty is highest.
All bases have recycling tank and children creche.
Energy reserves are primarily converted to minerals via production rushing. The conversion ratio is 2/1. Other energy reserves usage is rare and don’t impact overall faction power: subverting enemy bases and units, protection for subverting own bases, buying technologies, cornering global economy.
Economy, labs, and psych boosting facilities produce about same multiplicative effect at each point in time. I’ll use the term energy multiplier instead of individual economy, labs, psych ones for the sake of simplicity.
Psych allocation
Some effects can be converted to others by mean of modifying energy allocation. Whenever this allocation change needed I will always convert between economy and labs keeping psych amount unchanged. The main reason for it is that both economy and labs proportionally benefit faction power. The more of them you have the better. Whereas psych just quells drones. There is some sweet point of psych amount generated that quells just enough drones effectively. Anything below or above it is a waste.
Unequal allocation penalty
I assume most of the game you allocate more to labs than to economy. With this in mind any conversion from labs and economy has additional beneficial effect in dropping unequal allocation penalty. This effect is roughly about 3% of total energy loss per each 20% of labs-economy allocation difference per level of Inefficiency.
Normalization
I'll measure all effects comparing to INDUSTRY. This way INDUSTRY will have weight of 1.00 and other effects will be expressed in term of it.
Effect weight calculations
Industry
I’ll list a formula for Industry effect on production even though we take Industry effect as a base with constant value of 1. We’ll reference this formula below for other effect conversion purposes.
<minerals increase> = 0.1 * [<minerals production> - <support>]
Research
Labs excess received as a bonus of Research effect increase can be converted to extra production by directing it to economy for production rush. We also need to account for labs allocation change.
<labs excess> = <energy> * <labs allocation> * <labs multiplier> * 0.1 <energy increase due to labs excess conversion> = <labs excess> / <labs multiplier> <energy increase due to drop in labs allocation> = <energy> * 0.3 * Inefficiency * [<labs allocation> + <economy allocation>] * 0.1 * <labs allocation> <energy increase combined> = <energy increase due to labs excess conversion> + <energy increase due to drop in labs allocation> <reserves increase> = <energy increase> * <economy multiplier> <minerals increase> = <reserves increase> / 2 = [<energy> * <labs allocation> * <labs multiplier> * 0.1 / <labs multiplier> + <energy> * 0.3 * Inefficiency * [<labs allocation> + <economy allocation>] * 0.1 * <labs allocation>] * <economy multiplier> / 2 = <energy> * <labs allocation> * <economy multiplier> / 2 * [0.1 + 0.3 * Inefficiency * [<labs allocation> + <economy allocation>] * 0.1] = 0.05 * <energy> * <labs allocation> * <economy multiplier> * [1 + 0.3 * Inefficiency * [<labs allocation> + <economy allocation>]] <minerals increase> = 0.1 * 0.5 * <energy> * <labs allocation> * <economy multiplier> * [1 + 0.3 * Inefficiency * [<labs allocation> + <economy allocation>]]
Energy
This is a new effect introduced for the ease of further calculations and conversions. One step on Energy effect scale changes raw energy yield by 10%. It works similar to other proportional effects like Industry, Research, Growth. It doesn't exist in a game as described but is convenient to use in further calculations since many other effects can be expressed through it.
Let’s compare it with Industry and Research effects. We will keep psych intact and distribute extra energy to economy and labs proportionally to their current allocations. Total Energy weight will be a combination of economy and labs effects.
<economy increase> = <energy> * 0.1 * <economy allocation> / [<economy allocation> + <labs allocation>] * <economy multiplier> <minerals increase (economy)> = <economy increase> / 2 = <energy> * 0.1 * <economy allocation> / [<economy allocation> + <labs allocation>] * <economy multiplier> / 2 <labs increase> = <energy> * 0.1 * <labs allocation> / [<economy allocation> + <labs allocation>] * <labs multiplier> <minerals increase (labs)> = (<energy> * 0.1 * <labs allocation> / [<economy allocation> + <labs allocation>] * <labs multiplier>) / (<energy> * <labs allocation> * <labs multiplier>) * 0.5 * <energy> * <labs allocation> * <economy multiplier> * [1 + 0.3 * Inefficiency * [<labs allocation> + <economy allocation>]] = 0.1 * 0.5 * <energy> * <labs allocation> * <economy multiplier> * [1 + 0.3 * Inefficiency * [<labs allocation> + <economy allocation>]] / [<economy allocation> + <labs allocation>] <minerals increase> = <minerals increase (economy)> + <minerals increase (labs)> = <energy> * 0.1 * <economy allocation> / [<economy allocation> + <labs allocation>] * <economy multiplier> / 2 + 0.1 * 0.5 * <energy> * <labs allocation> * <economy multiplier> * [1 + 0.3 * Inefficiency * [<labs allocation> + <economy allocation>]] / [<economy allocation> + <labs allocation>] <minerals increase> = 0.1 * 0.5 * <energy> * <economy multiplier> * [1 + 0.3 * Inefficiency * <labs allocation>]
Growth
First of all I like to exclude population boom from Growth effect evaluation. Population boom is insanely powerful feature that completely breaks game and blows situational Growth effect value out of proportions. That’s why all social engineering mods designers try to make it harder to trigger by delaying positive Growth effect applications from society models. Population boom definitely adds to Growth effect value. I just don’t want to calculate it and leave its evaluation for subjective experts. In this article “Growth effect” means “Growth effect without population boom”.
Growth doesn’t contribute to your faction power immediately. Instead it speeds up population growth therefore increasing your future faction development potential. The analysis of exact impact is a little bit speculative but here is it. First, I emulated city population growth on a period from turn 1 to turn 200. I assumed city tile produces 3 nutrients with recycling tank and worker tile production increases from 2 to 3 nutrients during this period. Variation in above numbers don’t seem to change anything much anyway. Then I varied growth rate between -3 and +5 and evaluated how number of worked squares (population + 1 city tile) changes. Results look pretty consistent and showed about 7% worked squares increase per each level of growth rate on average. This may seem strange at first as one would expect exact 10% population increase with all equals but this how numbers added up. Maybe it is the nature of rounding when fractional changes do not manifest themselves in small amount. I don’t care much. Feel free to share your own emulation and results.
Second, we need to account for empire size growth. With 7% more extra workers you can build proportionally more extra settlers and bases. So, you’ll probably end up with 7% more bases at any given time.
Total population is the product of base count and average base size. With each one increased by 7% the product increases by 14%. I take that as the combined impact of Growth effect on total population.
Bigger population gives you proportionally bigger minerals and energy yield. So we can express Growth effect weight with Industry and Energy as below. Keep in mind that the Growth impact is stronger when it is applied earlier so it impacts longer future timeline. At the very end of the game it worth nothing barring population boom. That’s why formula below contain game left play time portion.
Growth = 1.4 * <left play time portion> * [Industry + Energy]
Economy
Economy effect is drastically not linear. We can try to estimate each type of Economy change on its own and then assign combination values to each step on Economy scale. Such Economy effect change types are: 1) energy per base, 2) energy per square, 3) commerce rating.
Economy (energy per base) = 1 / (0.1 * <energy>) * Energy Economy (energy per square) = <.base worked tiles> / (0.1 * <energy>) * Energy
Commerce rating gives you +1 per each faction you are not at war with and per each base participating in trade.
Economy (commerce rating) = <trading factions> / (0.1 * <energy>) * Energy
| from | to | energy per base | energy per square | commerce rating |
|---|---|---|---|---|
| -3 | -2 | 1 | ||
| -2 | -1 | 1 | ||
| -1 | 0 | |||
| 0 | +1 | 1 | ||
| +1 | +2 | 1 | ||
| +2 | +3 | 1 | 1 | |
| +3 | +4 | 2 | 1 | |
| +4 | +5 | 1 |
The exact Economy effect weight is calculated in the attached spreadsheet for each level change.
Efficiency
Efficiency effect is threefold. It decreases inefficiency, decreases number of b-drones, and decreases penalty for uneven economy/labs allocation. Let's consider them one by one.
Inefficiency change
Inefficiency is the amount of energy lost at base due to low Efficiency level. We need to calculate how this loss is changed with Efficiency level change. Inefficiency formulas will account for children creche (+1 Efficiency)
<inefficiency> = <energy> * <distance> / [64 – 8 * [4 – [EFFICIENCY + 1]]]] <inefficiency change> = <energy> * <distance> / 8 / [5 + EFFICIENCY] ^ 2 Efficiency (inefficiency) = <energy> * <distance> / 8 / [5 + EFFICIENCY] ^ 2 / (0.1 * <energy>) * Energy
b-drones changes
We'll calculate b-drones impact by comparing it to the amount of psych required to quell produced drones. Number of b-drones per base on normal size map and on highest difficulty.
<.b-drones per base> = <.base count> / (1.5 * [4 + EFFICIENCY]) – 1
If the above value is negative then there are no drones.
<.b-drones per base change> = <.base count> / 1.5 / ([4 + EFFICIENCY] ^ 2) <psych required> = <.b-drones per base> * 2 <energy required> = <psych required> / <psych multiplier> <energy required> = <.base count> / 1.5 / ([4 + EFFICIENCY] ^ 2) * 2 / <psych multiplier> Efficiency (b-drones) = <.base count> / 1.5 / ([4 + EFFICIENCY] ^ 2) * 2 / <psych multiplier> / (0.1 * <energy>) * Energy
Allocation penalty change
Finally, the penalty for unequal economy/labs allocation.
<energy penalty> = 0.03 * [4 – EFFICIENCY] * [<labs allocation> + <economy allocation>] * [<labs allocation> - <economy allocation>] / 0.2 <energy penalty difference> = 0.15 * [<labs allocation> + <economy allocation>] * [<labs allocation> - <economy allocation>] Efficiency (allocation penalty) = 0.15 * [<labs allocation> + <economy allocation>] * [<labs allocation> - <economy allocation>] / (0.1 * <energy>) * Energy
Support
Support translates to absolute minerals bonus/loss. Free minerals for new base is a small one time bonus and doesn’t contribute much to the effect value over the course of a game.
All levels of Support are quite achievable with vanilla SE models. In fact you can go as far as -5 to +4 even though rates beyond -4 to +3 give no additional benefits.
<minerals increase> = <freed minerals>
Police
Police drone inducing or quelling effect can be expressed through Energy same way as for b-drones. Police effect is twofold, though. The quelling part from level -2 to +3 works always. Whereas pacifism inducing part from level -5 to -2 manifest itself only when military units are outside of borders. For Police positive rates we also need to account for police unit support.
<less drones (Police positive)> = <Police positive rating transition effect> * <unit police power multiplier> <less drones (Police negative)> = <Police negative rating transition effect> * <how often explorer/attacker units are beyond borders> <psych saved> = <less drones> * 2 <energy saved> = <psych saved> / <psych multiplier> <energy saved> = <less drones> * 2 / <psych multiplier> Police = <less drones> * 2 / <psych multiplier> / (0.1 * <energy>) * Energy Police positive (support) = -1
Here are Police negative rating transition effect for reference.
<less drones from -5 to -4> = <number of explorer/attacker units> <less drones from -4 to -3> = 1 <less drones from -3 to -2> = <number of explorer/attacker units> - 1
All Police ratings are quite achievable. In fact, vanilla SE models give you -8 to +4 range. Though, values beyond -5 to +3 do not confer additional benefits. Later in the game Brood pit gives you +2 POLICE. So your future achievable rate range easily shifts to -6 to +6. So I assume all ratings can occur equally frequent except extremities which can occur slightly less frequent.
Morale
To evaluate Morale effect we will compare it to morale boosting facilities maintenance. This approach is, of course, highly speculative but so is everything in this article. We’ll use it as a first order of approximation. Morale extreme ratings are achievable but not that easy. So I give them less occurrence frequency in weighted mean calculation.
| from | to | morale attack change | morale defense change | occurrence frequency |
|---|---|---|---|---|
| -4 | -3 | +1 | +1 | 0.2 |
| -3 | -2 | +1 | +1 | 0.5 |
| -2 | -1 | 0 | 0 | 1.0 |
| -1 | 0 | +1 | +1 | 1.0 |
| 0 | +1 | +1 | +1 | 1.0 |
| +1 | +2 | 0 | +1 | 1.0 |
| +2 | +3 | +1 | +1 | 0.5 |
| +3 | +4 | +1 | 0 | 0.2 |
Weighted mean for morale per Morale level is 0.6 for attack and 0.8 for defense. I'll average it as 0.7.
Morale boosting facilities
Command center is the cheapest and most applicable morale facility at the beginning of the game as ground units are the majority of you army then. Later on focus shifts toward the air units and Aerospace complex. Then comes Bio-enhancement center for all units. I do not count Naval yard for naval units morale boost as they rarely need it.
| parameter | CC | CC + AC | CC + AC + BC |
|---|---|---|---|
| Total minerals cost | 40 | 120 | 220 |
| Total maintenance | 1 | 3 | 5 |
| Total moral levels | 2 | 2 | 4 |
| minerals cost per moral level per turn | 0.2 | 0.6 | 0.6 |
| maintenance per moral level per turn | 0.5 | 1.5 | 1.2 |
We will approximate progression in facilities minerals cost and maintenance as follow.
| parameter | early game | middle game | late game |
|---|---|---|---|
| minerals cost per moral level per turn | 0.2 | 0.4 | 0.6 |
| maintenance per moral level per turn | 0.5 | 1.0 | 1.5 |
Other considerations
Morale facilities are half as effective at Morale SE levels -4 to -2. Weighting this across all possible Morale ratings it gives about 0.9 correction coefficient for morale facilities effectiveness.
One doesn't need to build morale facilities in every base but in those mass producing units only. This essentially decreases their cost and maintenance per morale level raised. The multiplier is the portion of bases with such facilities.
Morale SE changes morale of all units at the right time. Whereas you need to maintain your facilities all the time to produce higher moral units. That is an additional multiplier to Morale weight which is the reciprocal to proportion of the time when you actively need it.
Combat mechanics
I believe no one would argue that combat advantage is crucial during the war time. Unfortunately, morale doesn’t proportionally contribute to combat advantage due to multi round combat mechanics. At Civilization I time casualties were exactly proportional to attack/defense strength ratio. Veteran unit with 50% bigger strength would then win proportionally more often. In other words, two veterans worth exactly three recruits. Civilization II and beyond screwed this proportion by introducing multi round battle and healing concepts. Now unit with 3:2 odds wins 100% of the time. With following healing it makes it invulnerable and army of them unstoppable. On top of that they guaranteed to reach elite status due to their indestructibility. Even Civilization I with balanced winning odds was conquest oriented. Whereas SMACSid Meier's Alpha Centauri just turns into a blitzkrieg every time after you got weapon advantage. It doesn’t matter whether you wipe out the whole planet with 10 needlejets or other faction does. It is still boring and doesn’t align with original game goal.
Back to the morale. Look at the picture in the attached spreadsheet on battle tab that depicts 10 rounds battle winning probability as a function of single round winning probability. Other authors researched battle probability function in SMAC and found it is goes even steeper then red line on a chart. Essentially 40% probability (2:3 odds) is a guaranteed loss, whereas 60% probability (3:2 odds) is a guaranteed win. With this in mind it is obvious that morale makes HUGE difference when your single round probability is around 40-60% range (with all other modifiers accounted for). Otherwise, when you are way out of this range, then morale gives you NOTHING! In other words morale helps tilting tie battles but becomes virtually useless for both sides in case of greater unit strength disparity. Besides, morale and modifiers boost is limited. Whereas weapon/armor progression is practically not. Morale still counts for psi and probe combat, though.
Summarizing the above I state that morale boost to your army competitiveness is critical in the early game but completely nullifies later. Therefore, I like to assign corrective coefficients to the Morale formula of 2,1,0 respectively for early, middle, late game stages.
Getting everything together
Per level of Morale effect rating.
<minerals saved (minerals cost)> = 0.7 * 0.2;0.4;0.6 <reserves saved (maintenance)> = 0.7 * 0.5;1.0;1.5 <energy saved (maintenance)> = <reserves saved (maintenance)> / <economy multiplier> <additional Morale coefficients> = 1 / 0.9 * <portion of bases with moral facilities> / <portion of war time> * 2;1;0 Morale = <additional Morale coefficients> * [<minerals saved (minerals cost)> / <INDUSTRY minerals bonus> * Industry + <energy saved (maintenance)> / (0.1 * <energy>) * Energy]
Where
- 0.7 is morale level per Morale effect rating.
- 0.2;0.4;0.6 is facilities mineral cost spread across 100 turns for early;middle;late game, correspondingly.
- 0.5;1.0;1.5 is facility maintenance for early;middle;late game, correspondingly.
- 0.9 is the correction for facilities morale increase ability due to halved modifier on low Morale effect ratings.
- <portion of bases with moral facilities> is what it is.
- <portion of war time> is the portion of time you wage active war (when morale is needed).
- 2;1;0 is battle mechanics adjustment coefficient for early;middle;late game, correspondingly.
Planet
Planet effect has following applications: improving psi combat odds, increasing worms capturing chances for positive ratings, impairing fungus square production for negative ratings, decreasing impact on global warming. Of these first one works during the whole course of the game as you keep fighting native life as well as using own psi units. Worm capturing is a nice addition to the army up to the middle of the game. Later on your produce worms faster than capture them. Impairing fungus production affects you only if you were using it in first place. It usually is marginal as only small part of your workers are placed on fungus. Besides, any fall in fungus production will trigger workers to switch to improved tiles instead thus reducing total fungus production impairment impact. Global warming happens only later in the game and can be easily offset by building specialized facilities.
I will calculate the Planet effect on a the psi combat same way as for Morale effect by comparing it to the cost and maintenance of life-cycle boosting facilities. Then I'll add a correction coefficient for natives capture chance.
Planet effect psi combat bonus is always 10% which corresponds to 10 / 12.5 = 0.8 of life-cycle level.
Life-cycle boosting facilities
In order of appearance: Biology lab, Centauri preserve, Bioenhancement center, Brood pit, Temple of Planet.
| parameter | BL | BL + CP | BL + CP + BC | BL + CP + BC + BP | BL + CP + BC + BP + TP |
|---|---|---|---|---|---|
| Total minerals cost | 60 | 160 | 260 | 340 | 540 |
| Total maintenance | 1 | 3 | 5 | 7 | 10 |
| Total lifecycle levels | 1 | 2 | 3 | 4 | 5 |
| minerals cost per lifecycle level per turn | 0.6 | 0.8 | 0.9 | 0.9 | 1.1 |
| maintenance per lifecycle level per turn | 1.0 | 1.5 | 1.7 | 1.8 | 2.0 |
We will approximate progression in facilities minerals cost and maintenance as follow.
| parameter | early game | middle game | late game |
|---|---|---|---|
| minerals cost per lifecycle level per turn | 0.6 | 0.8 | 1.0 |
| maintenance per lifecycle level per turn | 1.0 | 1.5 | 2.0 |
Other considerations
One doesn't need to build lifecycle facilities in every base but in those mass producing units only. This essentially decreases their cost and maintenance per morale level raised. The multiplier is the portion of bases with such facilities.
Morale SE changes morale of all units at the right time. Whereas you need to maintain your facilities all the time to produce higher moral units. That is an additional multiplier to Morale weight which is the reciprocal to proportion of the time when you actively need it. This portion of time is bigger than just war time since you benefit from Planet also while defending from randomly popping natives. Still war time is more important psi combat application so this portion won't be 100% but some number in between actual war time portion and 100%.
Combat mechanics
Same combat mechanics considerations apply to psi combat as to normal one. There is one important difference, though. Psi combat doesn't use weapon and armor values. Therefore, it depends on morale/lifecycle levels only. That is why I give lifecycle criticality factor 2;2;2 throughout the whole game as opposite to morale importance that is shaded by weapon-armor disparity later on.
Getting everything together
Per level of Planet effect rating.
<minerals saved (minerals cost)> = 0.8 * 0.6;0.8;1.0 <reserves saved (maintenance)> = 0.8 * 1.0;1.5;2.0 <energy saved (maintenance)> = <reserves saved (maintenance)> / <economy multiplier> <additional Planet coefficients> = <portion of bases with lifecycle facilities> / <portion of applicable time> * 2;2;2 Planet psi combat = <additional Planet coefficients> * [<minerals saved (minerals cost)> / <INDUSTRY minerals bonus> * Industry + <energy saved (maintenance)> / (0.1 * <energy>) * Energy]
Where
- 0.8 is lifecycle level equivalent per Planet effect rating.
- 0.6;0.8;1.0 is facilities mineral cost spread across 100 turns for early;middle;late game, correspondingly.
- 1.0;1.5;2.0 is facility maintenance for early;middle;late game, correspondingly.
- <portion of bases with lifecycle facilities> is what it is.
- <portion of applicable time> is the portion of time the Planet rating is critical. I.e. waging war with psi units as well as protecting bases against natives.
- 2;2;2 is battle mechanics adjustment coefficient for early;middle;late game, correspondingly.
Probe
This is IMHO the least useful effect of them all. Mostly because it makes difference so rare in the game and doesn't impact game outcome in general. Moreover, it takes only few probe teams on a front line to prevent enemy from capturing your cities and it is the only thing you need SE Probe for. The only notable exclusion is Miriam’s ability to set +3 Probe rating at the beginning of the game when she doesn't have cover ops centers yet. And even then it makes no difference for her when she loses technological advantage.
Overall I would rate SE Probe weight as zero but leaving this again for other player expert opinions.
Talent
Same calculation as for drone quelling.
<drones quelled per level> = 1 <psych saved> = <drones quelled per level> * 2 <energy saved> = <psych saved> / <psych multiplier> = 1 * 2 / <psych multiplier> Talent = 1 * 2 / <psych multiplier> / (0.1 * <energy>) * Energy
Summary tables
Effects
| Effect | early game | middle game | late game | average |
|---|---|---|---|---|
| Economy | 3.68 | 2.70 | 1.16 | 2.51 |
| Efficiency | 1.35 | 2.99 | 3.37 | 2.57 |
| Support | 1.33 | 0.67 | 0.22 | 0.74 |
| Morale | 1.53 | 0.32 | 0.00 | 0.62 |
| Police positive | 0.63 | 1.44 | 0.54 | 0.87 |
| Police negative | 1.02 | 0.86 | 0.36 | 0.75 |
| Police | 0.83 | 1.15 | 0.45 | 0.81 |
| Growth | 2.69 | 2.56 | 1.13 | 2.13 |
| Planet | 1.15 | 0.48 | 0.14 | 0.59 |
| Probe | 0.00 | 0.00 | 0.00 | 0.00 |
| Industry | 1.00 | 1.00 | 1.00 | 1.00 |
| Research | 0.77 | 1.22 | 0.82 | 0.94 |
| Talent | 3.07 | 1.28 | 0.36 | 1.57 |
Observations
Fixed unit count effects such as Economy (energy per base/square/trade), Support (minerals per unit), Police (drones per police), Talent(1 talent) have descent initial value and fade down the road comparing to percentage based effect like Industry as overall economy flourishes.
Weight wise most lucrative effects are Economy and Efficiency followed by Growth, Talent, Industry, Research, Police, Support, Morale, Planet.
Police positive weight fluctuates around same value because of double police unit ability activated later on in the game. Police negative weight shows negative effect of dumping Police rating to negative values while having military units outside of the borders. Without pacifism its negative impact can be ignored.
Growth fades toward the end of the game as expected due to due to game running out of future.
Both Morale and Planet psi combat weights decline due to increased availability of morale or life cycle boosting facilities and ease of their support with grown economy.
The only effect noticeably growing in value toward the end of the game is Efficiency. More or less stable effects across the course of the game are Industry, Research, Police positive. Other effects are constantly declining.
Models
See exact values in the attached file. On a high level, best models are Democracy, Knowledge Wealth, Cybernetic, Eudaimonic. Free Market is average on average but shines without pacifism and if it gets to the +2 Economy, obviously.
Overall calculated weights correlate pretty good with other players opinions I’ve heard on forums.
