Matrix Games Forums

Here is a list of the changes that I have added or
plan to add to the next version of "Hannibal," 1.0.4:

1) Correct the crash that can occur rarely when recruiting
in Cisalpine Gaul and Philip of Macedon is active and located
in that province.

2) Improve AI decision-making logic at the Normal and Hard
levels when deciding to deploy to the field. This corrects
the situation in which the AI didn't regard leaderless
Carthaginian armies as valid targets.

3) Correct error that allowed duplicate instances of the
"Two New Units With Hannibal" option card to be in the
player's hand. There should only be at most one of these
cards in the player's hand at any given time.

4) Improve logging especially at startup.

5) Correct certain unreported errors.

6) Enforce the rule that Elephants cannot be doubled as a
result of the Inspired Troops option from Hannibal's Punic
Tricks. The game was allowing this and it should not.

7) When the "Suprise" option from Hannibal's Punic Tricks
is in play the tiles will be marked as a reminder that all
units receive a special bonus for the first round of combat.

8) Support mouse wheel scrolling when a field is "open"
and which can be scrolled in the ordinary way.

9) Pontic Aid will be renamed to something more historically
accurate. The leading candidate is Illyrian Aid.

I'll update this list as needed. If anything seems to be missing
please post in this thread and I'll add it to the list if it
is appropriate.

Pontic Aid changed to Illyrian Aid?

Hi Mercenarius

Have been meaning to post about this for some time and then keep forgetting.

It is to do with Roman naval movement. Basically some Roman fleets seem to be moving extra spaces which is really anoying when i am trying to blockade a Roman Army in Spain.

Sequence of events:

1) Rome invades Spain and capture Emporoin and the stays in the city.
2) Carthagians besiege Emporoin
3) Carthagians moves their entire fleet into the Iberian sea to stop the Roman army slipping away.
4) The anoying bit, Roman gathers its fleet from Ionian sea and moves into the Tyrrhenian Sea picks up fleet units from their moves to the Med and then to the Iberian sea (once it also started in the bay of Terranium as well and then followed the above). From the way I am reading the rules Roman fleet units are moving to far. The Iberian and Tyrrhenian could individually get to the Iberian sea but not as a joint force and how the units from the Bay of Terranium got their i am not sure.

Pontic Aid changed to Illyrian Aid?

You have validated the existence of this thread already! And that's appropriate as it was your suggestion. [8D]

Talos Run5

Are you 100% certain that the AI didn't use an Extended Move option? Those are valid for naval movement and the AI will use them pretty aggressively at sea under circumstances like the one which you described.

What about a simple logistics system? Too time consuming for this next patch?

Wouldn't mind hearing a simple "yay" or "nay" on this: http://www.matrixgames.com/forums/tm.asp?m=2793143.

This "extra" long Roman naval move has happened to me as well. The Roman AI might have used up a card but I got no message of any card played and no opportunity to respond with a card.

Well, since we have a second confirmation, I should add the third. Imo, the computer does use the card, but the player is not notified.

When the AI uses its Extended Move option you don't get to respond the way you do to other options. This use by the AI is just like your use of Extended Move or Extended March (by Hannibal). If you put that card "On Tap" or use it when prompted the AI doesn't get a chance to play a response card.

There is a notice but it appears in the Force Display Area at the bottom of the screen. It's just one of the text messages. It doesn't stay up that long and most players probably don't notice it for that reason (especially when Verbose AI is turned off in the preferences).

I exempted the Extended Move (and Extended March) cards from the usual response pattern for a number of reasons. The biggest is that I didn't want the players to spend time trying to ruin the AI by playing cards like "Remove 1 Latin Allied" every time the AI wanted to use an Extended Move card.

Anyway, this is how it's supposed to work. I suppose that the manual could be a little clearer on this point.

Thanks for the response to my question. I shall look out for it next time i play and see if their is a message appearing, i have not noticed one but then i have not been looking in the force display area when Rome moves its fleets so could have easily missed it.

What about a simple logistics system? Too time consuming for this next patch?

Yes, because of the work getting the AI to know how to deal with it. [:(]

But I'll try to accelerate getting 1.0.4 into a public beta so that I can start thinking about 1.1 and the other things that we have talked about.

AI Improvement: when selecting cavalry losses, make AI always select Roman cavalry first and Latin later - this is something a player would always do, no reason to keep AI at disadvantage.

Another AI Improvement: smarter use of Siege cards. AI is taught to use them on Major cities, but it looks like it ignores the size of the garrison. E.g. AI attempts to retake Cis. Gaul. Genoa is defended well, but Placentia has only one unit and the minor towns two each. AI attacks Placentia and uses its only siege card to remove that one miserable defender. A smarter move would be to wait and use it on a minor town. Granted, maybe AI doesn't count that far ahead, but a check that the garrison has at least two units (and the attacking army size is larger that a certain constant (6-7?) would help it a lot (even if card is not used this turn, it could be more beneficial later).

AI Improvement: when selecting cavalry losses, make AI always select Roman cavalry first and Latin later - this is something a player would always do, no reason to keep AI at disadvantage.

It's true that the Roman cavalry are not worth as much as the allied cavalry but I wanted to make the AI impartial in choosing losses between Roman and Latin Allied units. This is not the best "game" choice for the AI to make but it's reasonably historical.

So that was my reasoning. However, you can make a case that the Romans actually leaned on their own cavalry first. So I'll consider this change. [:)]

Another AI Improvement: smarter use of Siege cards. AI is taught to use them on Major cities, but it looks like it ignores the size of the garrison.

That is a good point and I'll try to tweak this in the next release.

So that was my reasoning. However, you can make a case that the Romans actually leaned on their own cavalry first. So I'll consider this change.

Thanks for considering it I'd still stick to my own rationale (i.e. AI needs all the help you can give it without making it cheat) though.

That is a good point and I'll try to tweak this in the next release.

I think we are done with this thread than

I spoke too soon.

Another AI improvement. Rally Legion card use. AI seems to be programmed to randomly select a legion to rally, Latin or Roman, which usually doesn't matter (although a true Roman would probably sacrifice an allied legion first). However, there are some battles where it needs to learn to rally more selectively: battles at Roman gates. I just had a game where AI used this card and rallied Latin legion, leaving two Urban legions behind. Needless to say, a human player would never do this.

Now, there may be some wicked rationale behind this - Urban legions may be considered a garrison unit not worth rallying - I just want to make sure it's done on purpose.

Be careful with the AI improvement suggestions, nalivayko. Soon it will be impossible to beat the AI at hard difficulty. [:D]

Speaking seriously, I believe that AI doesn't always have to make the best choice in a combat situation. Many of the commanders are pretty mediocre (like they were in real life - the true geniuses are rare), so it only makes sense that they sometimes do stupid things, like rallying a wrong unit. I make mistakes all the time playing as Carthage. C'est la guerre.

Of course if somebody like Africanus is in that kind of a situation, the chances are he most likely makes the right choice. But I'd like to see the AI keep making mistakes, based on the commander's abilities, though maybe the case of urban legions is a bit too obvious, as they are such great units that one would expect even the stupidest of commanders to realize it and act accordingly.

In the future, it would be interesting to see the commander ratings have more of an impact on what can and can't be done during a battle. I've always liked wargames that give a bigger role to the characters and abilities of commanders, because even elite troops can suffer defeat if their leaders are incompetent, and sometimes a gifted commander can prevail against all the odds in an impossible situation. Maybe add some characteristics to the generals? Something like that is already in the game, like Hannibal's special cards, Fabius's ability to retreat before battle etc. But having more would be great. [:)]

I lose one game in five on Hard, usually rage quitting over some stupid move I made without thinking My biggest challenge nowdays is trying to beat AI under 10 moves, 5 being my record. 15 turns is my comfort zone (i.e. I am upset with myself if I drag the war any longer). Therefore, I doubt my suggestions will make AI impossible to defeat.

Speaking seriously as well, I disagree that AI's ability to make mistakes should depend on commander's rating. In theory, it works out fine. In practice, it gives you an unfair advantage over AI. You, as a player, are not handicapped when commanding either Hannibal or Hanno. Why should AI be forced to suffer where you are not forced to suffer? Besides, the use of some cards is already limited to generals with certain rank.

Another AI improvement. Rally Legion card use. AI seems to be programmed to randomly select a legion to rally, Latin or Roman, which usually doesn't matter (although a true Roman would probably sacrifice an allied legion first). However, there are some battles where it needs to learn to rally more selectively: battles at Roman gates. I just had a game where AI used this card and rallied Latin legion, leaving two Urban legions behind. Needless to say, a human player would never do this.

Yes, this is something which should already have been improved but it slipped between the cracks as it were. I'll improve the code for this at the higher levels in 1.0.4. It's not a huge change but I agree that we really notice it when an Urban Legion is available but isn't selected.