The AI is not the usual quality seen in other FFG products...

By MMOfan, in Star Wars: Outer Rim

First, I am a huge fan of FFG. As a traditional wargamer I am also heavy into thematic games FFG produces.

I think I have almost all these great adventure games like a complete collection of Lovecraft games (and dare I say most expansions too), a living room full of LotR LCG, Runebound expansions, XWing, and - of course - the App assist solo experiences in Descent2 and SW Imperial Assault...

So I was excited to see Outer Rim announced and - as a solo player - was very interested in buying it. Certainly after having played a 3 player game it seemed destined to succeed as a great solo game...

But ... after several solo sessions, the solo Bot as it NOW stands is subpar to many of my other solo games frankly. The Bot is not very consistent, it deletes text on ability cards and makes for some very poor play in around 40% of its actions.

The reason for this post is to try to convince the developpers to try correcting it in future expansions. It is not very interesting to see the AI Bot go/buy for cards it can not even use... then buy a job that it needs to go to the other side of an endless board and then waste money on things it has zero value for...

The game has such a HUGE potential, but sadly - compared to the other recent FFG products - its solo AI Bot is not up to its job.

I have one rule that keeps things fresh and objective when I play solo games:

If I start to support for the AI to win, it is not a good sign actually: the better solo experiences gives you a great fun challenge packaged in a compelling theme.

In the case of Outer Rim, the theme package is awesome (and even good in 2_-3 player) but I am always wanting the AI Bot to give me a better game, which it lacks at the moment.

So dear developpers: try to put out an expansion with a better AI.

Thanks in advance.

Do you think what you're looking for is an AI that more literally replicates the actions of a player?

I really enjoy the AI they created, but they certainly went for dynamic equivalence versus strict literal translation.

I appreciate their choice - I feel the AI moves like a player and takes cards like a player, so you can interact with them through the market decks much like a player. I feel like this was what they decided was most important to them, versus a more literal translation of what a player would do.

I think a more literal version could be made, but would it be worth the effort to play through it's turn? Probably not for me, but I'm sure for many!

Or is there something else you think is needed? Some missing fun factor or challenge factor? I personally have had a ton of fun with it so far, and I also know it's easy to give it a heads-start on points (as suggested in the rulebook) if I'm needing more of a challenge.

I really respect your opinion and it's clear you have great intentions with starting this discussion. I hope at least some fans could help make an AI you might better enjoy. Let's be honest, only fans have made the X-Wing AI, right?

By the way, by far what I dislike the most about the AI (probably the only thing, but it's not insignificant) is that it skips all of the text on the cards, especially that it skips the character's text. I would be very interested in specific AI versions of the characters that better differentiated them from each other. Again, I can see why they made this choice, but its truly sad not to see much difference at all in playing versus each different character.

I've been thinking a lot about what you've said about this topic over on BGG - I think the AI kind of works for what it is but also agree in the sense that at this point it is much more enjoyable with other players over playing a solo game.

I realized the rest of my reply was a derailment as I started talking about specific additions to home-brew a better AI system. I think I'll start a new thread for that and I hope you (both) will contribute.

Back on topic I will say I agree with you and would love to see FFG work on and release a more complex/dynamic AI for the game. I think that's beneficial in the long run and I hope they will consider this idea. I wouldn't be surprised if in an expansion they are planning more cards for the AI deck, but I think I'd like to see more than that for the AI.

Honestly, I don't think it's really fair to compare the AI from this to things like the App driven games like IA and Descent. IMO it's pretty clear that the AI in Outer Rim is an add on extra, and not a focus. I also feel there's probably not that much incentive for FFG to improve that part of the game, as I imagine the majority of players are interested in it for multiplayer.

Granted, I can understand someone being disappointed/angry if they bought the game assuming it would be more than it is.

I actually had more fun playing as two different characters than using the AI. I mean, the AI is fine and all but I find the fact that it can’t use bounty hunters is a bit limiting.

I think this is just a game that requires people to play with. And that's ok, that's just the nature of some games.

But if I was to try to create a solo experience for this game, I think I'd steer clear of having an AI Player and would instead try to create some sort of set of external events that altered the decks and board state in some ways. I'd have three other characters on the map and would given them each one gear, one mod, and one encounter token. They would each gain a Negative Rep with one random faction. They wouldn't have turns or move, but could be interacted with (they'll sell their gear or mod to the player for +1,000 over its cost) and they'll sell their crew for 5,000. They will defend their crew if it has a bounty and the solo player is trying to collect, but only if they'd roll more Combat dice than their Wanted Crew. Having these "NPC Players" stationary on the board would allow for some of the Player vs Player interaction (like the WANTED bounty), and does a little to simulate other players finding and revealing some of the board/gear during the course of the game. Then just, three times after the players' turn, randomly select a Market Deck and remove its top card from the game, using the newly revealed card to move patrols (move toward the middle of the rim if its on one of the two tiles closest to an and, otherwise coin-flip whether it moves right or left). After every third turn, randomly determine one patrol and destroy it.

Then, simply have the player simply race to some sort of goal before Turn X, which could be variable and even drawn from a goal deck to make each game feel especially different (maybe it's getting 10 Fame, Maybe it's getting 100,000 credits, maybe it's getting 6 Fame and 50,000 credits, maybe it's having a Positive Rebel Rep and destroying the Third Level of Imperial Patrol, etc. ...).

My gripe with the AI is that 8 out of the 10 cards allow the AI player to gain fame at the cost of 8000 credits when it cannot do a job (which usually results in gaining fame as well) as most of the time it will not occupy a space with a patrol that it will have negative reputation with. I have played three games against the AI and that is how it beats me every time (third game was an extended game as well).

On 6/28/2019 at 7:47 PM, CyberAdam64 said:

My gripe with the AI is that 8 out of the 10 cards allow the AI player to gain fame at the cost of 8000 credits when it cannot do a job (which usually results in gaining fame as well) as most of the time it will not occupy a space with a patrol that it will have negative reputation with. I have played three games against the AI and that is how it beats me every time (third game was an extended game as well).

You can turn your credits into fame, too, using the luxury deck. You can gain fame with personal goals and ship goals. You can gain fame from bounties and patrol-hunting. I know it can feel unfair on its face, but the AI’s credits-to-fame simulates all these fame opportunities that would be a pain to track or replicate for the AI.

You can slow the AI down, however. Leave expensive gear on the gear deck - 8,000 on gear that gives them low-to-no benefit? Yes, please. Attack the AI with a Wanted job or the Firespray to knock them down and make them spill some cash. Once you see your options, I think it feels a little more fair.

On 6/27/2019 at 2:15 PM, tafattel said:

I actually had more fun playing as two different characters than using the AI. I mean, the AI is fine and all but I find the fact that it can’t use bounty hunters is a bit limiting.

How was it playing two different characters? What about the secret cards? I’m pretty much a solo gamer but have been on the fence in picking this up.

It was quite fun actually. I’ve done it a few times now and enjoyed it. Secret cards aren’t really secret but I’ll usually just place them face down by that character’s area and only reference them on that character’s turn. That way I usually am not keeping the secret in mind and it doesn’t really make a difference in play.

When it comes to playing games with a solo mode I tend to just play 2+ player by myself as I make better decisions than the AI. For example, Star Saga has the Nexus AI for solo play, but when I played it drawing AI cards I kicked it's butt it made decisions that made no tactical sense, but when I played it as me being both players I kicked my own butt 🤨 as in, the Nexus player won due to me playing it as I would.

I tend to do this alot with games I am playing solo meant for 2+ players. I can even play X-Wing solo and still win, lmao.

I only got Outer Rim today after watching loads of game play videos and already encountered a query, to which I was trying to find an answer for. Which actually blends in with this topic nicely.

I am playing as Han versus AI Lando, as previously pointed out above I can see that the AI in it's current state is going to make for a dull game if all it is doing is buying cargo, it doesn't test for anything and therefore the crew cards will be redundant. So the AI card told me to Revel a Contact Token and gain it as a crew for free, it turned out to be Garindan, databank card #10, as the AI does not benefit from any of it's abilites apart from adding any to it's ship or personal stats. So what is the benefit of having this crew member? It feels like all it is doing is keeping Garindan available for a bounty by me the player? Which just so happens to be the next one in the market.

Not having played a full game yet I can't fully commit to what I think I want to do, but looking at the Data Bank cards there are and all teh AI cards. But to enable the AI to be a bounty hunter I think it it draws a crew card that is any of the following;

Faction : Hutt

Faction : Syndicate

Text Specifically says it is a bounty hunter, i.e; 4-LOM

Then when it comes to drawing an AI card if I am told to purchase something I will look at the one it tells me and compare it to the bounty deck and choose either the closest destination or the cheapest one or the one that gives the highest reward, I'll have to test these as I am new to teh game.

Would be nice to know if any of you think this would not work?

Also I can't see anything regarding if a crew appears for AI but hasn't got any benefit, as the AI doesn't test technically Lando should pay 2000 creds for Garindan, Garindans ability is to SECRETLY look at a Contact Token, I can't see how this is benefitting Lando. So basically I am discarding it rather than taking it.

Edited by T9FURY
Added an update