Fixing SMACX Bugs

[scient]

However, I thought it's a bit odd that the scrambling unit ("Tachyon Tactical" of the Drones) starts its flight from the north pole when it's displayed. Also, the message displayed reads "... near Iota Sector", but Iota Sector is on the other side of the planet. The Tachyon Tactical actually sits in "Echo Bay", a Caretaker base a few tiles to the SE.

It seems that the position of the scrambling unit isn't set properly. I don't know whether you regard it as a problem - the crash *is* fixed, so I'll rather be grateful for the fix than nitpicking about details. However, if you *do* want to fix the location of the scrambling unit too, then I'll gladly run some further tests.

Ya, I thought that was odd as well. I'll have to look at a case where scrambling doesn't crash vanilla game and then values being reset. And no, please do nitpick. I won't get offended that's the whole reason I want folks to test these out to add polish of things I might have missed.

Also, is there anything specific which you're currently working on that I can help with? Creating a test save for a specific problem, analyzing a specific bug, that sort of thing?

Well, I've been focusing on probe and probe related activities to sort them once and for all. So far this has mainly revolved around enhanced probes not being able to mind control high SE morale factions and how survival/success probabilities are calculated. While working on these I've glazed over some of other probe actions but if you can find or think of any others I'll have a look. The function is too big to go through them all. You can just barely make out small "selection" box at tip of pointer in pic below.

[Psyringe]

It seems the manual should have said or if their Probe rating was +2 rather than +4 since the manual also says that at +3 probe bases and units cannot be subverted (p. 161).

Yeah, I think the manual is inconsistent with regard to high probe bonuses. While the info in the manual *can* be a good indication of what the designers wanted, we nevertheless shouldn't be oblivious to the fact that the manual can have "bugs" as well. Also, there's a chance that the manual is simply outdated in parts and that the designers consciously changed a game mechanic after the manual was already printed.

I never realized that there was a bonus for former or colony pod. This seems rather strange. Wouldn't it make more sense that civilians would be easier to subvert? Maybe it should be if unit is former or colony pod, divide nWorth by 2.

From a "realism" perspective that makes sense. From a "gameplay" perspective however, I think the rule makes more sense as it currently is. In post #220, scient outlines how nWorth is calculated - we don't know which factors are involved, but scient suspects "unit morale, armor and special abilities" to play a role. These are all factors that colony pods won't have, and formers will only have to a minor degree. Hence, colony pods might end up severely undervalued with regard to their nWorth, and the designers might have decided to double their nWorth to compensate for that. From a gameplay perspective, a colony pod is much more valuable than a garrison, so the subversion cost should reflect that (imho). (Note: I haven't tested whether it currently does that, so we're both working on assumptions.)

[vyeh]

You can also "attack" with any combat unit a treatied probe unit in your territory and have the option of questioning the probe unit and releasing it. In such case, the probe unit is returned to the nearest base of the treaty partner and there are no consequences.

The full quote (p. 8 of the PDF on the Alien Crossfire disk) is

"Algorithmic Enhancement
Used by Probe Teams only. Probe Teams equipped with this ability have their chance of failure cut in half when acting against normal targets (a 60-40 action would become an 80-20 action), and have half the normal chance of success penetrating the defenses of factions with the Hunter-Seeker Algorithm, or whose social engineering choices have rendered them "immune" to mind control. Becomes available upon discovery of Nanominiaturization (B8)."

So:

In my opinion, the algorithmic enhancement is not meant to affect the cost (where you have opponent SE >= 3, the cost algorithm uses the value of SE = 2 instead of returning an immune message), but only meant to change the chance of success (halves the failure attacking normal targets -- no HSA and SE < 3; and for attacking HSA or SE >= 3, it nullifies HSA or makes SE = 2 and it halves the chances of success under those circumstances).

So the nWorth code should be as you specified with the following change:

>=3: if no enhancement, return immune message
if enhancement, nWorth * 2

To answer your question in post #220, I would not move SE morale one step down that ladder. What I would do is put a check for enhancement in the section of the code that looks for >= 3 and if there is an enhancement, use nWorth*2; otherwise return the immune message.

Psyringe,

While the manual has its problems, the datalinks, the manual and Prima's strategy guide are objective evidence. Speculation about the designers changing a game mechanic after the manual was printed or that the manual is "outdated" (there was a small manual errata on the Alien Crossfire disc) shouldn't negate the usefulness of the manual as a tool in this project.

Concerning colony pods and formers, I don't have an interest in subverting them. My interest is in subverting units that I can't build or that are a threat. For that matter, I am usually not worried about a former or a colony pod being subverted because they aren't a threat to me and they can be replaced. I just found it strange that there was an adjustment for them.

I agree with scient that we should leave the adjustment as is.

You may have missed it in an earlier post, but I believe that scient first wants to fix bugs and then later consider improvements.

Your post brought up something that you could do to help scient with the probe function. You said that scient suspects "unit morale, armor and special abilities" to play a role in calculating nWorth.

With the scenario editor, how about seeing what is the cost to subvert units? I'd suggest starting with a scout and changing its morale level. Then you could fix the morale level and try various levels of armor.

You could also compare the cost to subvert a unit to its mineral cost.

[Petek]

I agree. Is it not the same bug, then?

Agreed.

Try right-clicking and loading other queue slots- if there's nothing there, or only a colony pod, we've had a fail.

I don't understand what you mean by "loading other queue slots". I opened the base screen and took various actions on the open queue slots. Never saw the option to add another hovertank former to the queue.

Say what? It's year one.

I was trying to say that perhaps I needed to have the save from the previous game in which hovertank formers could be built.

Still can't reproduce this bug from my own games.

Petek

[Darsnan]

One other bug before I'm outa town: In the #TERRAIN section of the alpha(x).txt, if the "Raise Land" and "Lower Land" options are set to disabled and the the Weather Paradigm is disabled, the AIs' are still able to raise and lower land. Note that I am pretty sure I have seen this in the past (I tried to get the AIs to do it last night, but to no avail), and that under the circumstances I listed above the AIs should not be able to raise or lower land. If anyone can confirm this for me I'd appreciate it, otherwise I'll continue playtesting to see if I can get the AIs' to replicate this.

scient: I'd say don't bother with this until someone else can confirm it, and then you can handle it as you wish.

D

[Psyringe]

While the manual has its problems, the datalinks, the manual and Prima's strategy guide are objective evidence. Speculation about the designers changing a game mechanic after the manual was printed or that the manual is "outdated" (there was a small manual errata on the Alien Crossfire disc) shouldn't negate the usefulness of the manual as a tool in this project.

Yes, that was my point. The manual is a useful tool (and, in absence of actual contact with the designers, one of the best sources of information that we have), I just used the "enhanced probes" example to point out that it's not infallible.

You may have missed it in an earlier post, but I believe that scient first wants to fix bugs and then later consider improvements.

I'm aware of that (I actually did read through the whole thread before I posted ). That's why I wasn't arguing on a "it should/shouldn't be fixed" basis in my post - as long as you don't claim that the current behavior is a bug (which you didn't), there's little point in arguing about whether scient should fix it since it automatically becomes a low priority issue anyway if it doesn't classify as a bug. I saw your statement as a comment on game design ("Shouldn't civilian units be easier to subvert then military?"), so I explained why (imho) the current design might make more sense. (I see your point though.)

Your post brought up something that you could do to help scient with the probe function. You said that scient suspects "unit morale, armor and special abilities" to play a role in calculating nWorth.

With the scenario editor, how about seeing what is the cost to subvert units?

Great minds think alike, I'd say. That's exactly what I've been doing for the past hour - creating a testbed to assess subversion cost. It's going slow since I have never used the scenario editor before, but I should have something ready by tomorrow evening.

[Buster Crabbe's Uncle]

I don't understand what you mean by "loading other queue slots". I opened the base screen and took various actions on the open queue slots. Never saw the option to add another hovertank former to the queue.

When the new game begins, they're unlabled. Slot 1 is the former, slot 2 a sea former, slot 3 hab complex, ect. It sounds like it's working for you the way it does for me, so I assume they should be there. [You don't even see the new unit loaded until you mouseover something or click on the resource screen.]

I was trying to say that perhaps I needed to have the save from the previous game in which hovertank formers could be built.

I doubt it. I think it's probably some .ini hold-over effect. But once I've saved the new game, the build queue is there when I resume playing the next day. [Or after I've been playing a different game of AC.]

Still can't reproduce this bug from my own games.

Beats me. I produced it with v.2, so it's not that. I wonder if you did reproduce it, but haven't loaded the unlabled queues...

[scient]

Oki, for enhanced probes I'll cap SE morale at 2 if it's >= 3. That will be in next update, very easy to do. Hopefully I'll have analysis on success/survival probability done soon.

[BlackCat]

Looking for other things, I stumbeld over this thread - don't know if there are a loop or it just is the result of udescisiveness by the AI.

http://apolyton.net/forums/showthread.php?t=158549

[vyeh]

I saw your statement as a comment on game design ("Shouldn't civilian units be easier to subvert then military?"), so I explained why (imho) the current design might make more sense. (I see your point though.)

Actually, more surprised that such a factor would be included. I'm wondering if the cost of subversion enters into an AI choosing to subvert a unit or not. Perhaps the AI was subverting too many colony pods and formers in testing, so this was put in to focus the AI on military units.

Great minds think alike, I'd say. That's exactly what I've been doing for the past hour - creating a testbed to assess subversion cost. It's going slow since I have never used the scenario editor before, but I should have something ready by tomorrow evening.

According to the manual, the defending (subvertee) unit's probe SE should be a factor in cost.

According to the manual, polymorphic encryption on the defender is supposed to double the cost of probe action.

According to the manual, algorithmic enhancement affects the probability of success (there is no mention of cost).

There are certain techs that give the probe unit a greater chance of success: polymorphic software, pre-sentient algorithms, digital sentience, self-aware machines, and mind/machine interface (alphax.txt only mentions success rate).

I would imagine that the defending unit's cost in mineral rows is probably a factor (as opposed to depending on armor or special abilites).

The manual (p. 135) says:

"The cost of mind control operation against a base depends on a number of factors, including the base's size, the enemy's energy reserves, facilities and special operations in a base, and whether or not drone riots are occuring."

In looking at morale, there is a (-) factor when a drone riot is occurring in that unit's home base. This may have an effect on the cost of subverting a unit.

Children's Creche is suppose to reduce a base's vulnerability to mind control (p. 166). Genejack Factory is suppose to increase a base's vulnerability to mind control. HQ stops mind control.

[scient]

According to the manual, the defending (subvertee) unit's probe SE should be a factor in cost.

The AI uses the same code so this could be true.

According to the manual, polymorphic encryption on the defender is supposed to double the cost of probe action.

Check.

According to the manual, algorithmic enhancement affects the probability of success (there is no mention of cost).

I haven't posted it yet but check. It also had modifier depending on if unit has enhancement and defender has HSA.

There are certain techs that give the probe unit a greater chance of success: polymorphic software, pre-sentient algorithms, digital sentience, self-aware machines, and mind/machine interface (alphax.txt only mentions success rate).

I thought these just increased probes unit morale maxing them out to elite? There is no code in the area that generates success/survival rate that indicates checking these techs. There are further down the random value generation which I can check but I don't think these are factors. In data links as well as when I was skimming manual all I could find was +1 to probe's base unit morale.

I would imagine that the defending unit's cost in mineral rows is probably a factor (as opposed to depending on armor or special abilities).

This would be part of nWorth generation.

The manual (p. 135) says:

"The cost of mind control operation against a base depends on a number of factors, including the base's size, the enemy's energy reserves, facilities and special operations in a base, and whether or not drone riots are occurring."

In looking at morale, there is a (-) factor when a drone riot is occurring in that unit's home base. This may have an effect on the cost of subverting a unit.

One of the parameter's passed to probe success/survival probability function is the state of city (normal, riot, stapled, maybe others?). This value is fixed for units.

Children's Creche is suppose to reduce a base's vulnerability to mind control (p. 166). Genejack Factory is suppose to increase a base's vulnerability to mind control. HQ stops mind control.

Genejack and Covert Op Base are used for nWorth generation. Covert Op Base is also used inside success/survival code. HQ is working correctly. I have seen no code regarding CC but I'll keep an eye out. Maybe it modifies city state parameter.

[vyeh]

The AI uses the same code so this could be true.

What do you mean by same code? I was referring to manual , p. 161, which defines the various probe SE values.

I thought these just increased probes unit morale maxing them out to elite? There is no code in the area that generates success/survival rate that indicates checking these techs. There are further down the random value generation which I can check but I don't think these are factors. In data links as well as when I was skimming manual all I could find was +1 to probe's base unit morale.

The techs I mentioned may work by increasing the morale of probe units. I believe when you discover these techs existing probe teams as well as newly built probe teams get a +1 morale boost.

One of the parameter's passed to probe success/survival probability function is the state of city (normal, riot, stapled, maybe others?). This value is fixed for units.

Are you saying that stapled affects the chances of mind control?

And are you saying that it doesn't affect mind control whether or not unit's home base is undergoing drone riots?

Genejack and Covert Op Base are used for nWorth generation. Covert Op Base is also used inside success/survival code. HQ is working correctly. I have seen no code regarding CC but I'll keep an eye out. Maybe it modifies city state parameter.

This is certainly a place that Psyringe can test empirically. We can see if the presence of a children's creche affects cost (nWorth) or success/survival.

[scient]

What do you mean by same code? I was referring to manual , p. 161, which defines the various probe SE values.

The same modifiers used for players is used by ai.

Are you saying that stapled affects the chances of mind control?

And are you saying that it doesn't affect mind control whether or not unit's home base is undergoing drone riots?

Stapling does yes, it even tells ya when you do it (only success/survival). I dunno about unit's home rioting causing negative effect. It does have an effect if you try to mind control a rioting city.

[BlackCat]

I accidentally hit the wrong button for a transport foil with one movement point and popped a pod - it was hit by a tidal wave and got all it's movement points back.

Not exactly a bug that you want to exploit - the chances for an IoD is probably higher

[vyeh]

The same thing happens when a land unit pops a dimensional gate. It moves to a new location with full movement point.