User:Spike

Hi, my name is Spike. I live in London. My mates all played XCom when it came out, when they were feckless students, but I had a job so I didn't get to play. I'm making up for lost time now. I love retro tactical SF games. I like to play Laser Squad, MegaTraveller, any old rubbish.

I think one of the reasons is these TU-based tactical games are all variants of the old miniatures rules by GDW - Snapshot, Mayday, Azanti High Lighting etc. In the day, how we wished for a computer program to help us with the tedium of playing those games by hand. Even Laser Squad would have blown us away.

Having said that I don't play any head-to-head stuff, like X-Com 2000. I imagine I would get my arse kicked.

This site is fantastic to use, so it's nice to be able to make some small contributions to the site, and the game.

Musings

To avoid spamming the site Discussion pages, and spamming people with 'watch' enabled on those pages, I'm going to start doing the decent thing and composing my musings here on my own page. If I get my thoughts edited and reasonably coherent I will then transfer them to Discussion pages and then onto actual articles.

Here are some of the topics I am interested in at the moment...

Economics

The Geoscape game is a classic resource game and delivers lots of great game play in its own right, even when not intercepting or doing tactical missions. Economics is key. Looking at the the economic articles on the site at the moment, they focus on making money. What I am interesting in is 'fixing' the economics so it all makes more sense.

Efficient market

At the moment it is possible to sell lots of things for either profit or loss that does not make sense. I would like to fix the buy and sell prices so that they balance out and there are no egregious opportunities for arbitrage, or the reverse. For example I've calculated that based on the maximum profit you can make out of Elerium and Alloys (the raw materials in the economy of XCom), the price of Alloys and Elerium should both be quite a lot higher. Or alternatively some of the prices of manufactured products should be lower.

I would like to just true up all the prices of everything to reflect the economic costs of production, and to allow them to be sold at a modest profit. (And bought at a modest premium, see below). It's not good that there are so many 'black hole' items that are hideously unprofitable to produce, and a few 'optimum' items that everyone manufactures all the time.

See Also

Until I incorporate them here, see also my sections in Talk:Manufacturing_Profitability

Capital costs and interest

Correct pricing needs to take account not only of raw materials, labour and maintenance inputs, but also capital costs (initial hiring, facility building). For this an interest rate (cost of capital) needs to be posited for XCom. Given XCom's status as a covert internationally funded organisation, reasonable access to money markets could be assumed. For game purposes a capital cost of 5% - 10% is probably fine. Even 1% per month would probably work fine given the game's short time horizons.

Actually the return to capital can probably be calculated quite easily (based on production costs vs profits on the most profitable items). If X-COM were actually to be allowed to borrow, it should be charged at least near to this much interest or otherwise a free money machine is created. Because of X-COM's short-term, high stakes objectives and potentially massive future cash flow, it would probably borrow very heavily if it could to build up a strong position. But then, who would lend to an entity that not only doesn't exist, but is fighting a life or death struggle against powerful forces beyond our comprehension?

A tricky question though with capital costs is how to amortise them over the production of multiple items. Obviously the worst case is to charge the entire capital cost to the first unit produced (a possibility worth considering under 'External Markets - Buying', below). We could make some educated guesses about the order of magnitude, based on the likely market size. Again, see below under 'Inelastic Demand' for some guesstimates - thousands for small arms, hundreds for aircraft weapons, dozens for aircraft. Unfortunately this is a better 'efficiency of scale' than X-COM gets in the standard game, so it would make the game easier. For purposes of game balance over realism, we need to amortise over numbers that make sense in the time frame of a single X-COM campaign. For balance, that might be more like ten for small arms, 2-4 for craft weapons and 1 for aircraft. This comes in to play more with the 'External Markets' angle.

External market

At the moment you can buy and sell conventional equipment but only sell some advanced equipment (and not all of that), and not buy any of it.

Selling

As long as egregious profits are not being made (see preceding section), I think you should be able to sell all manufactured products eg aircraft. After all, the funding nations (and others) would no doubt like to get their hands on aircraft with such advanced characteristics. The problem of course is setting a fair sale price. Probably the game designers didn't have the time to do the exhaustive manufacturing analysis that has since been done on this site and other fan sites. With the benefit of this data, we can set fair prices for all manufactured components and allow them to be sold (by editing PURCHASE.DAT). Standard (or higher?) fixed profit ratios would apply for buy vs. sell pricing of manufactured goods.

Inelastic demand

Demand for advanced, alien and conventional items is totally inelastic in the standard game: the price does not drop no matter how many units you sell. On reflection, I don't think this is that unrealistic. For conventional items, XCom purchases and sales are a microscopic part of the whole traded economy in such items - outfitting a maximum of 250-odd soldiers will not make a dent. For advanced and alien items, demand is never really going to drop. Once the US special forces have been outfitted with Heavy Plasmas, the Israeli, Russian, and French special forces will want to catch up. (And the British SAS might get around to buying some Laser Rifles, ha ha).

Much less would the nations of the world say no to purchasing high performance hybrid aircraft. They would pay millions (and perhaps this is why the game designers did not permit the sale of aircraft).

I can see buyer fatigue setting in if hundreds or thousands of laser cannon are produced for sale, as these are limited to combat aircraft and there are only a few thousand front line combat aircraft in service at any given time. Another reason to diversify the options away from a handful of 'optimum' manufacturing items.

If you wanted a cap you could probably specify about $1 billion of inelastic demand for advanced items of each type. 7000 Heavy plasmas - about enough to equip 12 special forces battalions. A hundred or so of each aircraft type. After that the market might start to get a bit soft. And maybe the market for e.g. Mind Probes would be more limited.

I think there's a case that the very first few examples of each item would have a much higher price. But perhaps we can posit that as part of its funding agreement, XCom has agreed to sell recovered weapons etc back to the Funding Council nation at reasonable and stable prices. Possibly the sales are allocated to funding nations by rota in the early stages - that would make sense. The funding nations get allocated artefacts "by lots" and are then free to resell them to each other or, eventually, on the open market.

Buying

Slightly more interesting is allowing the 'External Market', in the form of military-industrial firms with links to the funding nations, to produce advanced items. This would then allow XCom to simply buy, rather than manufacture, items such as Avengers.

The pricing would simply be the total economic cost, as determined by the 'Efficient Market' process above, plus some profit margin - probably the standard buy/sell profit margin which is about 33%.

Now obviously you don't want to make advanced items available from the start, that would defeat a lot of the challenge of the game. From a game play point of view what you want to do is give the player a reasonable alternative of buying vs manufacturing the item, once they have done the necessary research.

Options for doing this... say that an item can be bought on the open market:

1. Once the item has been captured.

2. Once one (or more?) examples of the item have been sold onto the market.

3. Once the item has been researched by XCom.

4. Once one example has been manufactured by XCom (this could be hard to check for though).

5. Any of the above, plus a time delay. For 1-2 this would be related to the Research time.

In cases 1-2 the Research is being done outside of XCom which really changes the game. In this case there should be a big premium payable to the external arms company that owns the patents etc. This can be done in PURCHASE.DAT as the buy and sell prices can be set independently of each other, so set the Buy price at 200% of Sell price (vs the more normal 133% of sell price). Or at the very least we could put the price of the item up by say 50-100%. Keep in mind that as part of 'Efficient Pricing' we have (in theory!) already factored in the cost of the research effort (scientists, labs, capital etc).

In general though I don't favour bypassing the Research tree so that favours options 3-5. 3 is easy to check for in the game files. 4 is tricky, at least for alien items, because while you can see items under construction, you might miss the moment they are constructed. For non-alien items, the mere existence of the item in the game proves it has been manufactured.

As an extra variant you might say that only hybrid items, not pure alien items, can be manufactured by the external firms. So while laser rifles, a power suit or an Avenger can be built, and indeed mass produced, by a high tech arms company, only X-COMs "Deep Black" science labs can painstakingly assemble small numbers of genuine alien items like Mind Probes, Blaster Launchers, etc.

I'd be quite comfortable with this variant (hybrid only) and it makes it easier to police option 4 which is the most challenging and probably the best for gameplay. Or you could mix 3 and 4, charging a 50%-100% premium on purchases until such time as X-COM has produced their own example of an item.

There is also the question of delivery times. Delivery times for the Avenger are very long and compressing that to 72 hrs would seriously affect the game balance. I think that PURCHASE.DAT can be hacked to lengthen the delivery times for specific items. Probably the delivery time should be based on a reasonable setup such as a 100-space workshop with as many engineers will fit. We don't really want to provide advantages to buying the Avenger, we just want it to be a not unreasonable option.

One final question is availability of resources, especially Elerium which is the only thing that can't be manufactured. The price of Alloys can be computed with a floor equal to the labour & capital costs of production, and a ceiling based on the most profitable item you can produce with Alloys (Mind Probes I think from memory). The price of Elerium should be at least the value derived from the most profitable item you can build with Elerium. Potentially it is much higher, since Elerium is scarce and has other very valuable irreplaceable functions (hybrid aircraft fuel, winning tactical missions, saving the world...).

A strict interpretation would be track the amount of Elerium sold (and anything else that can't yet be manufactured by the external market at that time, eg Alloys / UFO Power Sources / UFO Navigations). You would then not be allowed to buy more manufactured items than could be built using the total resources sold up to date to the 'external market'. But apart from patching the game I can't think of any way to do this kind of double accounting. Actually I can, in XComUtil you could record all UFO recoverables after each mission, and then subtract from that all XCom stores and the inputs to all existing manufactured items (ignoring anything destroyed - presume comprehensive recycling of such valuable scrap). But that would be a hassle to code up.

There is an argument for a highly elastic price for Elerium, based on the total supply. Of course, as players we 'know' that (apart from variant self-imposed rules like 1-Mission or No Detection etc) there is ample Elerium just around the corner. That's the reality of the standard game, so in a standard game a fixed price for Elerium is probably justified.

The Elerium Standard

This brings me back to the beginning in a logical circle. Prices are all relative, so what is the reference point? You can do this one of two ways. Either take the most profitable item in the game, and work back from that to deduce the correct price for Elerium (pricing all other components along the way, eg Alloys, Power, Navigation). Or, you can take the price of Elerium as it is, and work your way 'up' the chain, correcting the prices for everything built from Elerium. I like the first method because it has the least impact on existing revenue-generating options from manufacturing. The revenue a player can create should stay exactly the same (proportionate to the inputs of labour, capital, raw materials); they just have many more ways of generating that revenue.

Detection

Detection by Aircraft

See my sections in Talk:UFO_Detection#Detection

Fixing Multiple Radar

See my section in Talk:UFO_Detection#Multiple_Radar_Effectiveness_Algorithm_and_Hack

Seb76 now has a brilliant fix for this bug in his UFO Extender. Also see the Tools section below for the Base Fixer.

The algorithm I used for the Base Fixer was:

smallf = (0.9) ** nsmallradar

largef = (0.8) ** nlargeradar

xdetshort = int (round ((1 - smallf * largef ) * 100))

xdetlong = int (round ((1 - largef ) * 100))

#special case backward compatible

if nsmallradar == 1 and nlargeradar == 1 :

xdetshort = 30

xdetlong = 20

I've confirmed TFTD also sufffers from a "Sonar Stacking" bug and is in fact identical to UFO:EU, the only difference being a different offset in BASE.DAT.

Phantom Radar Bug (Talk)

Is listed at ExploitsA#Phantom_Radar_Trick. Thanks for adding, though. Also, I'm under the impression that building a new structure, even starting one, eliminated Phantom Radar; so upgrading-in-place cannot be done. Arrow Quivershaft 21:52, 17 March 2008 (PDT)

OK thanks for that. i will test but you are probably right! removed this text:

This bug allows you to "upgrade in place", for example building a new Hyper-Wave Decoder over the top of an existing Large Radar, and retaining the detection capability of the Large Radar until the Hyper-Wave Decoder completes building. (Unless something else completes building first).

Spike 02:10, 18 March 2008 (PDT)

For XCOM CE (since I stare at this in a hex editor a lot):

* Completing any new structure causes the radar stats to be recalculated at the base the structure completes at.

* Starting new structures is perfectly safe.

-- Zaimoni 7:49 18 March 2008 (CDT)

OK I've done some tests and confirmed what Zaimoni said is right. Upgrade in place is possible. If you dismantle a small radar and build a large radar over it, the SR detection value applies until the large radar (or any other facility) completes building in that base. I see the detection values in base.dat and I also get detections/intercepts from the base that has "no radar". The Base Information screen detection strength shows zero but base.dat shows 10%.

Reinstating the "upgrade in place" comments in the Known Bugs entry!

Spike 17:47, 19 March 2008 (PDT)

Tools

Data Tables and Spreadsheets

Tactical Firepower

• X-COM (EU and TFTD) Battlescape Firepower - Spreadsheet comparing "close range/shock firepower" and "long range/skirmish firepower" of Battlescape weapons - it covers XCom and Alien hand-held weapons, XCom HWP weapons, and Alien Terrorist built-in attacks. Includes XComUtil variant weapons. The results are specific to a Target Type (Muton, X-COM Power Suit, etc) and you can set the attacker's combat skill levels. Also includes "lethality" calculations such as the probability of a first-shot-kill, the average number of hits to get a kill, and the average amount of TUs to get a kill.
• Includes some info for TFTD weapons, not yet including SWS weapons or Alien built-in attacks. The necessary information on TFTD Alien attacks is hard to find. I still need to merge in the TFTD Alien stats from the draft TFTD version of my Firepower spreadsheet.

Tactical Firepower Model

• Formula for Weapon Rankings

In a nutshell, I figure out the average armour-adjusted damage per direct hit (average of min + max, adjusted for % of range that is zero), then multiply that by the average # of direct hits per 100% TUs (assuming Firing Accuracy = 50). I then divide the target Health by the resulting value, to obtain the "%TUs per kill".

The "Weapon Ranking' number given is for whichever firing mode (auto, snap or aimed) gives the best value. I should probably list what the best mode is!

If your soldiers have a better average FA, it all scales linearly. So if you have FA=100 just cut (improve) the values by half. I picked FA=50 as its a typical starting value, and it was particularly the starting weapons I was interested in.

I tried lots of different "figures of merit" for weapons but I like this one best.

All HE is considered a direct hit on the Alien. For 4-square targets I add the reduced, armour-adjusted damage on the other 3 squares.

I did not bother with Incendiary as the damage model for incendiary is almost totally independent of the weapon used (the only weapon based factor is whether you get the target inside the area of effect or not, and how many times).

The very low (good) average values, such as for Heavy Plasma, will be very volatile around the average. The high (bad) numbers will be more consistent.

This Figure of Merit probably works best when the number is in the range 25% - 75%. Below that level, there is high volatilty and also I may not have accounted for overkill (minimum penetrating damage > Health) properly. Above that level, I have not accounted for reload time, nor "turn rounding errors" - such as: you can't burst-fire an AC 2.5 times a turn, you can only burst-fire it twice.

Another thing to note about this metric is that it is accuracy-weighted, so it assumes a non-trivial "firing problem". For situations such as point-blank where accuracy is not an issue, the rankings will be very different. I'll work on a different metric for point-blank. My Firepower spreadsheet, which generated these rankings, allows you to toggle the "turn rounding" on or off. With "turn rounding" off, the calculations are "instantaneous" measures of lethality, which make sense for intra-turn decisions, or when a mix of fire and movement is important (as it usually is). With "turn rounding" on, you get a better assessment of "continuous" firepower - e.g. when firing continuously over multiple turns.

It would also be good to have others check my assumptions in the spreadsheet!

Spike 22:22, 27 February 2009 (CST)

Near Miss Modelling

One thing I don't calculate is the effects of any area effect/HE near misses. So the figures I give should be taken as a minimum. In practice, HE effectiveness will be somewhat higher due to the extra damage from near misses - but by how much? It depends on a lot of factors, some I don't know and some that are highly variable (density of terrain / map objects etc).

Calculating the range and average damage done by any given near miss is complex but not impossible. There are a large number of calculations, one for every possible 'near miss' square with different GZ+X distance, different armour facings, etc.

What is truly difficult is estimating the absolute and relative frequency of rounds landing in all these possible 'near miss' squares. With sound knowledge of the error angles (vertical as well as lateral) generated by the game engine, it might be possible to estimate a probability distribution, for each possible range from launcher to target - but only for a flat, featureless plain. It's the terrain features - fixed and mobile - that introduce the biggest uncertainty about where a stray round will detonate. This is almost possible to model. Without modelling for collisions with map objects, the 'flat plain' model of near-misses will predict a worst case (minimal) damage level from near-misses. (Assuming that in the typical case, the accurate path from firer to target is relatively unobstructed. When this is not the case, the 'flat plain' scenario might actually predict better results than obtain in reality.)

One more problem with modelling Near Misses: the benefit of lucky near misses - whatever its value - is inversely proportional to overall accuracy. A perfect shooter gets exactly zero benefit from near misses. Which raises an interesting possibility. Maybe the "near miss benefit" could be estimated using a repeated experiments with a logger and a shooter with Firing Accuracy=0. Any damage done to the target would be due to a lucky near miss. This would be the other limiting case - the opposite of the perfect shooter. Then extrapolate between those two extremes to find a "near miss benefit" for that weapon as some function of {TA, range].

Other Tables

Data table comparing the "firepower", "payload" and other obscure characteristics of aircraft weapons: Aircraft Firepower Table

Also I have uploaded some spreadsheets concerning XcomUtil manufacturing profitability; efficiency of aircraft as radar detection platforms; effect of my proposed multiple radar fix on detection probability.

I have a version of the Manufacturing Profitability spreadsheet for TFTD, which I will put on the main pages when I can figure out how to wikify it. Also, this version calculates full Technician/Engineer costs, including fixed as well as variable costs, and calculates the payback/breakeven period.

Base Fixer

I've written a Python script that

• fixes the Paying for Dirt bug
• fixes the Phantom Radar bug
• fixes the Radar Stacking bug and implements my algorithm for improved effectiveness of multiple radars

It's working reasonably well now so I've uploaded the BaseFixer utility and documentation to this site. You will also need to install Python for it to work. I have also used XcomUtil's hook facility to integrate BaseFixer into XcomUtil, so it runs automatically as you play (updates whenever you switch from tactical battlescape to the geoscape view). Spike 12:00, 24 March 2008 (PDT)

TU Efficiency

Analysis and fixes (game file patches) relating to the fact that Aimed fire is less efficient than Snap fire. See Accuracy vs TU Efficiency.

Incendiary Research

Large Units On Fire

Here's another interesting test for setting Large Units on fire. Set AC-I power down from 42 to 1. At this IC power, there is no area effect. You can't even start a fire on one tile. All you can do is set units on fire with a direct hit. This means there is no way the other 3 squares of the Large Unit are in the IC area of effect.

Result: Whether you hit a one square regular unit or a 4-square large unit, if it goes on fire, all squares of the unit still go on fire. Along with the fact that the inflicted IN damage level vs Large Units is basically normal (see above), this tends to confirm that being on fire is a "unitref" property and, unlike HE and Stun Bombs, is not applied (multiplied) vs the number of squares the unit occupies. (In fact it would be good to check that HE is actually applied separately to each square, with different HE strength, potentially different armour facing, etc. Maybe the engine just simplifies things and multiplies damage by 4?)

Spike 07:56, 15 March 2009 (EDT)

Unit Burn Time

Unit Burn Time seems to be considerably cyclic rather than random. Very often, it just steps down by one from the previous value. I suspect this is because the "impact" routine calls the "end of turn" incendiary/smoke routine (see User talk:Seb76#Incendiary Bug). Sometimes it does not happen, maybe because some other factor has caused the Burn Time to increase.

Here's some data. This was done against a Reaper with AC-I power set to 1 (and 10% TU Snap fire cost). But I've seen similar results with standard AC-I.

Hit	H/H	Dmg/Avg	Burn Time
1	137/150	13	2
2	126/150 11	1
3	104/150 22	0	Very high damage (=13x1.7). Unit not on fire when game restored.
4	 96/150  8	7
5	 81/150	15	6
6	 66/150 15	5
4,5,6	 76/150	/9.33	5	(Auto) 9-10 (=5-6x1.7) is near minimum dmg
7	 56	10	4
8	 48	 8	4	What happened here? didn't cycle. MIN dmg (5x1.7).
9	 35	13	6
10	 25	10	5
11	 10	15	6	New shooter takes over
12	  2	 8	5

Spike 07:56, 15 March 2009 (EDT)

You're right on spot once again ;-) When a unit takes a direct hit from an IN round, it is set on fire (the duration is randomly chosen between 0 and incendiarydamagemodifier/20). Then the "end of turn" routine is called, which inflicts damage *and* decreases by one the number of turns to be on fire (done for all units). Then it sets on fire units standing in blaze. The duration is calculated as for direct hits; if the unit was already on fire, it'll take the maximum value between the new value calculated and the current value affected to the unit on fire. Not sure it is readable but I think you'll sort it out ;-) Seb76 10:51, 15 March 2009 (EDT)

Work in Progress

Underway

• Craft Firepower table for TFTD

To Do List

• Incendiary Research
• Firing Accuracy Research
• Seb76 UFOExtender
  • Verify that extra Smoke still blocks LoS
  • Verify extra Fire has its normal effects.
  • Check if Funky Fire/Smoke bugs still "work" with extra Smoke/Fire.
  • Use "View All Locations" to empirically check the effect of (fixed) multiple radars - does having greater %detection at the same range really increase UFO detections, or will the same UFO get deteced eventually anyway?
• Gameplay
  • Play an "Aliens Own Earth" game
  • Play a "Funding Only" game
  • Do a Cydonia mission - I've never stuck it out that far. Then do the mission again, but with conventional weapons and no Psi.
• Firepower Tables
  • Add all missing TFTD target and weapons data, in line with EU targets and weapons
• Economics
  • It would be interesting (and easy) to calculate the cost incurred when damage is sustained by aircraft. It would then be possible to put an economic value on the stand-off capability of weapons.
  • Efficient Market - Rational Costs
    • Understand impact of Research costs (capital and expense) on Manufacturing Profitability, and thus on "rational costs"
      • Necessary first step: Create a tree structure with the fully inclusive research costs to reach each node in the tree. If everything were profitable, it might make sense to attribute only the current nodes costs to any given item. Maybe it makes sense to amortise all research costs incurred since the last profitable item on the current branch? Or to share out the total costs along all paths, weighted by the profitability of each of the profitable nodes?
    • Adjusted buy & sell prices that better reflect the actual economic costs of manufactured items (assuming there would be demand for them outside X-COM). Then put this in a modified PURCHASE.DAT. The purpose of this would be to better harmonise manufacturing profitability of various item types.
    • Also make it possible to buy (rather than build) the manufacture-only item types. Calculate appropriate prices based on fully inclusive costs, normal buy-sell spread, and some estimation of capital costs including Research costs. (See discussion in Economics - External Market, above.)
    • Re-calculate manufacturing profitability, and manufacturing costs, based on Opportunity cost of diverting manufacturing away from profitable Laser Cannon.

Apocalypse Blog

Apocalypse: The Verdict

I got bored of this game very quickly. The real time combat is interesting to mess around with but ultimately a bit frustrating. I don't particularly like RTS's anyway, and this is not a great one. But it's the strategic game play that really falls flat. There is no story arc as in X-COM 1, I just didn't care about this artifical city and the weird alien invasion. Maybe it would've been better if the game was just about warring megacorporations without the alien "sub-plot". Then they could've built in some more plot.

If you took out the aliens and added in some more structured scenarios, it would work as a good remake of Laser Squad - the weapon mix, tactical focus and time frame seem to be similar.

Snags and Hassles

Like everyone I found it impossible to figure out how to load soldiers into a vehicle for a mission. Eventually I had to download and RTFM. I still don't know how to persuade the soldiers to take a vehicle home from a mission, so they walk home. Maybe I need to click on the building, or vehicle assignment screen, right away before they walk out of it, and reassign them to the vehicle before they start walking? Anyway, for now the walk will do them good. Good exercise and they can walk off all that aggression.

It seems like you automatically strip equipment off dead enemies, but you don't automatically strip the equipment off your own casualties, even if you win the battlefield. I am sure this is true because I am not only loosing armour, I am loosing Mind Benders. That's a hassle, if I have to manually take equipment off my dead guys, and make sure I do it before the mission ends. Maybe the equipment is just being blown up by HE? I noticed that the enemy seemed to be surprisingly vigorous in attacking the dead bodies of some of my guys, so maybe that's it.

It was totally non-obvious to me that you should raid neutral organisations for profit and plunder. I could've played for years without doing that, had I not read about it in strategy guides etc. It seems very cynical, not the vibe I would expect.

I don't like the Dual Wield stuff. Lots of games have dual wield, and I think it's always nonsense. Twin cannons, twin rifles? Give me a break. It is catering to the Arnie / Ninja crowd. If you use a 2 hand weapon in each hand you should get a big penalty on the primary and an even bigger penalty on the secondary. It would only ever be viable if you had no need for speed or accuracy and just wanted pure volume of fire. I can just about buy it with handguns, but not with anything else. Even handguns are supposed to be used with two hands, except for snap fire. If you think dual wield is accurate, ask yourself why no military, police or special forces unit anywhere, ever, uses this technique. It's only from dumb movies. I am annoyed that the mechanics of Apoc make dual wield advantageous. Apart from being stupid, it's an unfair advantage over the enemy since they don't use it (or do they?). I suppose I could just be self disciplined and just not use any weapons in dual wield. Stick to my principles!

I guess dual wield makes it harder to use grenades. Although apparently there is no TU penalty in Real Time mode for shouldering your rifle to grab a grenade? I find it too fiddly though. I need to learn, as grenades are powerful and they give an edge to the enemy if I can't use them.

Weapon Mixes

At the start it looks like the priorities are:

• Improve Agents' experience
• Conserve money
• Conserve ammunition

It looks to me like the Sniper rifle and M4000 have about the same firepower (accuracy x fire rate x damage), but the M4000 has much higher ammo consumption (both per unit of time and per hit). So the top choice would be the Sniper rifle. As there are not that many around at start, I would give the Sniper rifles to the people with the lowest Accuracy score (contrary to the USG advice), and then M4000s to the rest. As I capture more Rifles I would put the M4000s into stores (for the next batch of recruits).

Of course the Plasma gun has greater firepower but it's so rare and the ammo so limited and expensive at the beginning that I would only issue it to the Agents with the highest accuracy, and then only during a UFO recovery or alien capture, not during a regular raid. Train up combat skills using cheaper and more plentiful weapons, and save the Plasma gun ammo for the critical missions.

The Autocannon I would only issue to a minimum number of agents with the highest accuracy and strength, and use it as a support weapon. Probably to your starting Androids. One or two Agents with autocannon is fine. Also I would not bother with the AP rounds. Because of the low accuracy and low rate of fire, the AP firepower is lower than the other starting weapons. I would stick to using the area munitions (HE/IN), employing the autocannon as a support weapon. The one exception for AP would be against armoured targets - a very specialised mission in the early game.

Actually, according to my theoretical calculations, in Dual Wield the highest firepower (apart from the rare and expensive Plasma gun) is actually a pair of Lawpistols. This is because the lack of one-handed penalty pushes their firepower ahead of the M4000 and Laser rifle. Of course the ammo capacity is the worst, and the range is not good, and at close range maybe accuracy is less important.

Anyway, if we have the Marsec M4000, where's the Marsec Auto-Gun? That was always superior to the M4000, in Laser Squad anyway.

As I don't really understand the mechanics, the above may be totally wrong. Maybe some kind of TU factor gives a disadvantage to the Laser Rifle. For example, in Real Time, maybe a fast moving target won't be acquired in time by a slow weapon like a sniper rifle, even in snap fire mode. I think the Sniper rifle is the best for experience gain but I could be wrong about that too. The USG says experience gain is per shot. This is almost certainly wrong - surely it's per hit - but if it's right, the M4000 would be the experience-meister, since it spews out (inaccurate) rounds.

In terms of purchases, I found I pretty much buy all the personal arms on the market (apart from some of the armour and AP grenades). I'll probably find a use for all of it. Definitely buy all the Mind Benders and Laser Rifles as these are relatively rare - especially the Mind Benders.

Psionics and Psi Raids

There is not a lot of info out there on the net that I can find about the Psi mechanics and Psi training mechanics. The USG has an illegible formula that I can't decipher, and it's not clear if it applies to field/combat experience or just to Psi-Lab training.

I'm trying to figure out the Psionics by experiment, and in pursuit of this I've invented a variant of the Stun Raid, the Psi Raid.

As with a Stun Raid you can go on a Psi Raid and keep all weapons locked, and just use Psi, and you don't get shot at. Even if you come off the borders of the map and walk in and around the building. You get 0 mission points but you don't trigger any Organisational hostility from the organisation you are raiding. I was doing successful Stuns (well I got the hammer on the head icon, but no one actually fell over). I was doing successful Scans. None of it provoked any gang guards to attack with weapons. I did get one Psi attack on me. I don't have anyone who can attempt Mind Control, but I should also do some Panics to see if that provokes hostility. Maybe if I actually knocked someone out with Psi, that would provoke hostility. I have 4 Agents capable of Stun and one capable of Scan only, but even piling up on one target I could not actually make them pass out from a Stun attack. Still they are all only Rookies.

My intention was to train up the Psi skills from combat experience, but I did not see any noticeable improvements after combat. Certainly any improvements were less than 1.0 stats. Maybe you need to make Kills to gain experience. Maybe it can only be trained in the Base, not in combat. (In X-Com 1 and 2, raising Psi stats in combat creates a runaway feedback loop that is unbalancing, and also makes Psi Labs pointless apart from for screening, so maybe they fixed that in X-Com 3)

What would a successful stun look like? Does it decrease Health with a white bar like X-Com 1 and 2 (apparently not), or decrease Psi Strength until you go unconscious? Maybe use will use Scan in conjunction with Stun Gas to see how Stun Gas works.

It seems like not only can you not do a psi attack if you are not in Line of Sight (fair enough), but also not unless you are on the same map level as the target. So even if you are one square in front of a target and up one level, you can't use Psi on them. Maybe something else was happening. Two of my Agents seemed to start getting 0% all the time, even if they were on the same level, in LoS, without having moved, full TUs and full Psi Strength. Is there a maximum number of attacks you can make in one mission, or a maximum amount of Psi strength you can expend? Do the Mind Benders actually need to recharge back at base at some point?

- OK I think this was just due to targeting stronger opponents (higher Psi defence) and the probability, particularly of Stun, dropped to 0%. You definitely can do a Psi attack from a different level. And I don't think you ever run out of Psi attacks during a mission, having done a lot more attacks now. Also, Panic attacks also don't trigger hostility. But then, I still haven't managed to do a Stun or Panic attack that actually seemed to affect the target at all. The attacks are "successful" but don't seem to do anything to the target, or reduce any of its stats (as viewed via Probe). Hmm.

Also (not a Psi note but a general one), switching weapons to and from any area takes no TUs. The only thing that takes TUs is picking up. Dropping is free. So the backpack is no different from a holster. Swapping your autocannon into your backpack and pulling a grenade takes 0 TUs. Presumably the effect is similar in Real Time mode. Not sure I really like that, it seems sloppy on the part of the programmers.