Forum discussions are now taking place on Discord. For all account questions and concerns, please continue to contact Customer Support directly.

Keep updated on Pirate101 on Twitter @Pirate101, Facebook, Discord, and @KI_Alerts!

 Search  Updated Topics  Hottest Topics  Rules

Everybody is targeting Fan Flanders

AuthorMessage
Ensign
Dec 27, 2014
2
I've only bought crowns once, so I am not very lucrative in this game. The most powerful boss I have battled so far was Chok a' Kab (did I spell that right?), and I have fought him and other bosses more than once.

Every time I fight them it seems they target Fan Flanders. Don't get me wrong, he is a great companion but he dies quickly once he gets several hits onto him. I try to move or heal, but it's like the enemies are glued to him.

Does anyone else experience enemies targeting a certain companion?
Developer
As a Swashbuckler, Fan Flanders has both a very high damage output AND very low health.

The AI (quite correctly) determines that the odds of defeating you are better if Fan Flanders is taken out quickly.
Pirate Overlord
Mar 10, 2009
6204
Exemplary on Feb 2, 2015 wrote:
I've only bought crowns once, so I am not very lucrative in this game. The most powerful boss I have battled so far was Chok a' Kab (did I spell that right?), and I have fought him and other bosses more than once.

Every time I fight them it seems they target Fan Flanders. Don't get me wrong, he is a great companion but he dies quickly once he gets several hits onto him. I try to move or heal, but it's like the enemies are glued to him.

Does anyone else experience enemies targeting a certain companion?
As Ratbeard is so fond of saying: "Protect your squishies". Keep Fan back behind someone like Ratbeard that has Hold The Line talent. Fan can shoot from across the board so keep her well back and protect her. Any kind of shielding buff will only help too.
Captain
May 16, 2011
552
Exemplary on Feb 2, 2015 wrote:
I've only bought crowns once, so I am not very lucrative in this game. The most powerful boss I have battled so far was Chok a' Kab (did I spell that right?), and I have fought him and other bosses more than once.

Every time I fight them it seems they target Fan Flanders. Don't get me wrong, he is a great companion but he dies quickly once he gets several hits onto him. I try to move or heal, but it's like the enemies are glued to him.

Does anyone else experience enemies targeting a certain companion?
Every companion/character generates a certain level of threat. It's like this: say companion X has very low health and threat. Companion Y has very high health and threat. Enemies A, B, and C are ganging up on Y. As soon as X comes after them, A, B, and possibly C will start attacking X. C may stay behind and finish off Y, but may not.

Now, Buccaneers and Swashbucklers tend to generate the highest levels of threat. Witchdoctors and Musketeers are behind them, and Privateers are at the back. But that's generally for characters. For companions, it's: Swashbuckler/Buccaneer, Privateer/Musketeer, then Witchdoctors. I believe. But besides that, Fan Flanders tends to generate lots of threat, as she's a Swashie. If you want her to survive, I recommend you send her as far away from enemies as you can before she dies.

Just a tip.
Pirate Overlord
Mar 16, 2012
10631
Exemplary on Feb 2, 2015 wrote:
I've only bought crowns once, so I am not very lucrative in this game. The most powerful boss I have battled so far was Chok a' Kab (did I spell that right?), and I have fought him and other bosses more than once.

Every time I fight them it seems they target Fan Flanders. Don't get me wrong, he is a great companion but he dies quickly once he gets several hits onto him. I try to move or heal, but it's like the enemies are glued to him.

Does anyone else experience enemies targeting a certain companion?
The AI in the game will always target either the greatest threat or the weakest ( with the lowest health ).
BTW, Fan is a girl.
Dread Pirate
Jun 17, 2013
2743
Ratbeard on Feb 2, 2015 wrote:
As a Swashbuckler, Fan Flanders has both a very high damage output AND very low health.

The AI (quite correctly) determines that the odds of defeating you are better if Fan Flanders is taken out quickly.
Hellooooo Ratbeard!

I just had a quick question(s) about what the AI does when you use same class, very similar crewmembers in battle. Like say all Swashbucklers with around the same damage, same "other" attributes, pretty much the same armor rating, dodge (and other stats) and also all possessing equally low health.

What is the next step or factor that the AI uses to calculate it's best odds of defeating us, if 2 or 3 crewmembers are, stat-wise, more or less equal? Does the AI then move onto powers/buffs that the crewmember possesses to determine the "biggest threat"? Or is it simply a matter (nothing simple about it, I'm sure) of a programming algorithm that processes all of the crewmembers stat/attribute numbers and, based upon that calculation and resulting number, determines whom to attack?

And I thank ya!
Developer
ValkoorTheVictorio... on Feb 3, 2015 wrote:
Hellooooo Ratbeard!

I just had a quick question(s) about what the AI does when you use same class, very similar crewmembers in battle. Like say all Swashbucklers with around the same damage, same "other" attributes, pretty much the same armor rating, dodge (and other stats) and also all possessing equally low health.

What is the next step or factor that the AI uses to calculate it's best odds of defeating us, if 2 or 3 crewmembers are, stat-wise, more or less equal? Does the AI then move onto powers/buffs that the crewmember possesses to determine the "biggest threat"? Or is it simply a matter (nothing simple about it, I'm sure) of a programming algorithm that processes all of the crewmembers stat/attribute numbers and, based upon that calculation and resulting number, determines whom to attack?

And I thank ya!
The AI assesses all of its possible moves, then immediately throws out the bottom (worst) half of those moves. It then picks randomly from among the remaining moves. We have the ability to make some enemies "smarter" to choose moves closer to the top of the list. Some enemies choose on a bell curve so that they're more likely to choose something from the middle of the list than they are to pick the best choice or the worst remaining choice.

"Best Move" uses some complicated math to ultimately make a pretty simple assessment-- how's my team doing? how's their team doing? What's the best thing I can do to improve my team and degrade their team?

There are a LOT of things that we could still do to improve the AI and make it much smarter (I hate their options when they clearly don't know what to do) but experience has shown that the AI is plenty smart enough to make the game challenging enough for the widest part of our audience.
Dread Pirate
Jun 17, 2013
2743
Ratbeard on Feb 4, 2015 wrote:
The AI assesses all of its possible moves, then immediately throws out the bottom (worst) half of those moves. It then picks randomly from among the remaining moves. We have the ability to make some enemies "smarter" to choose moves closer to the top of the list. Some enemies choose on a bell curve so that they're more likely to choose something from the middle of the list than they are to pick the best choice or the worst remaining choice.

"Best Move" uses some complicated math to ultimately make a pretty simple assessment-- how's my team doing? how's their team doing? What's the best thing I can do to improve my team and degrade their team?

There are a LOT of things that we could still do to improve the AI and make it much smarter (I hate their options when they clearly don't know what to do) but experience has shown that the AI is plenty smart enough to make the game challenging enough for the widest part of our audience.
I appreciate the reply and the insight, RB. I also would like to thank you for explaining it in a way that is easy to understand, even for an algebraically challenged individual as myself (well, I get by...but I won't be programming any video games anytime soon). I'm sure that in a roomful of tech-jargon speaking programmers, my eyes would go crossed and I'd be making motor boat sounds with my mouth.

It's still pretty fascinating to me how video game programmers can make the AI "think" on different terms and at various levels. I know that is what you guys and gals do, but I've really come to appreciate all of the work, calculating, assessing and thinking, done by the actual humans, that goes into the programming, that in turn makes the program and AI appear to be thinking on a human, true enemy and creature level.

I do have some follow up questions about summoned minions and how the enemy AI responds and assesses the minions in battle in comparison to pirate crews. When a summon (or even pet) shows up in battle, does the AI follow the same "thought process" when assessing those minions, as in the AI still attempting to determine the whole, "overall how's my team doing? how's their team doing?" and "What's the best thing I can do to improve my team and degrade their team?", scenario? Does the AI view summons differently than "regular" crewmember threats, meaning that-- is there a separate algorithm or assessment that just correlates to summons/pets (but is still part of the overall assessing)?

Being that, most times, the summons don't possess as much of a threat as a pirate and or companion, and that they possess much less damage dealing capabilities and health, does the AI still treat them as just a part of the crew? Or do they get higher/lower priority when the AI is assessing it's next move? I know that sometimes the AI seems to focus on a minion when it pops up, when this happens is this more because the AI realizes that it can quickly eradicate it from the board or because the AI has deemed that eliminating it just best falls in line with the my team/their team aspect?

Oh and I absolutely do agree that the AI in P101 is clearly smart and potent enough, even given their lack of options at times "when they clearly don't know what to do". Sometimes they don't make the best decisions based on what's available, but hey neither do us humans (the similarities are pretty neat, actually). If I may add, what other options would you like to see them have when they aren't sure what to do?

I understand if you can't answer all of this, as I'm sure you all are right on the verge of releasing some new worlds and content. I'm just still very intrigued by all of this programming mumbo-jumbo, so I have to admit that my questions are a bit selfish in nature. But I've stated it before and I'll state it again-- I do appreciate you folks at KI taking the time to answer our questions and give us a glimpse of what you all do and just how you do it.

Thanks again.
Developer
They assess pets and summons with the same A.I.

Pets and summons come onto the board with a built-in "handicap" (the more summons, the greater that handicap) and the AI does not do as good a job thinking through that handicap as I would prefer. It knows that it's almost always the best choice to reduce the enemy's numbers, and it knows that the pet has less health, so it will tend to seek them out.

It's not so great at assessing whether the pet's damage handicap is enough to make it bounce off its armor or resistance. The AI does perform a few "simulated" combat calculations but there's only so deep that each calculation can go before the server performance hit outweighs the benefits of smarter A.I.

In an ideal world I'd have the server calculate every decision in a "virtual round" with 100% fidelity to all the game rules, and then decide what worked the best. One step up from that would be something like Big Blue that calculates each move, then predicts your next move, then all of its possible responses, then all of the responses to that, and so on down the line until it has calculated not just the effects of each possible move on the next round, but all the way through the entire match.

But as we all know, it's just one small step from that to Skynet, self-awareness, and the extinction of humankind, so we thought better of it.

I'm sure you'll agree the A.I. is plenty good enough for the vast majority of decisions.
Petty Officer
Feb 27, 2009
94
Chrissy Th'Blesser on Feb 2, 2015 wrote:
As Ratbeard is so fond of saying: "Protect your squishies". Keep Fan back behind someone like Ratbeard that has Hold The Line talent. Fan can shoot from across the board so keep her well back and protect her. Any kind of shielding buff will only help too.
But Chok' Akab comes before you get Ratbeard. Chok is on the quest for Ratbeard's ghost crew.

Quentin Davenport, lvl 65

Marcus Davenport, lvl 25
Dread Pirate
Jun 17, 2013
2743
Ratbeard on Feb 5, 2015 wrote:
They assess pets and summons with the same A.I.

Pets and summons come onto the board with a built-in "handicap" (the more summons, the greater that handicap) and the AI does not do as good a job thinking through that handicap as I would prefer. It knows that it's almost always the best choice to reduce the enemy's numbers, and it knows that the pet has less health, so it will tend to seek them out.

It's not so great at assessing whether the pet's damage handicap is enough to make it bounce off its armor or resistance. The AI does perform a few "simulated" combat calculations but there's only so deep that each calculation can go before the server performance hit outweighs the benefits of smarter A.I.

In an ideal world I'd have the server calculate every decision in a "virtual round" with 100% fidelity to all the game rules, and then decide what worked the best. One step up from that would be something like Big Blue that calculates each move, then predicts your next move, then all of its possible responses, then all of the responses to that, and so on down the line until it has calculated not just the effects of each possible move on the next round, but all the way through the entire match.

But as we all know, it's just one small step from that to Skynet, self-awareness, and the extinction of humankind, so we thought better of it.

I'm sure you'll agree the A.I. is plenty good enough for the vast majority of decisions.
What he say?

That is some pretty fascinating info Ratbeard and thanks for taking the time to share it. Actually, and quite surprisingly, I did understand it all, although I admit that it took a few calculations of my own before my brain could fully process it.

It's great to "hear" how all of this programming magic happens and to learn just what it means and why the enemy does what it does. I know that it is a bunch of code, algorithms and calculations but actually equating those things with the enemies we know and love (and hate), really adds a new, enjoyable dimension to the game...well, at least for me.

I understand what you are getting at in regards to the server performance vs. smarter AI benefits-- even though I don't completely understand. And yes, I certainly do agree that the Spiral AI is good and smart enough and in many cases almost too smart-- a job well done, for sure. I'm always impressed with some of the moves the enemies make in battle.

So, is this Big Blue of any relation to Deep Blue (distant cousin perhaps)? I gathered from a great many of your posts that you are quite the chess enthusiast and admirer...and I can certainly see as to how having a passion and deep respect for chess would go hand-in-hand with AI programming and program development. I don't know if any of us pirates would be ready for something like what Deep Blue could do. Heck, I often have a hard enough time thinking of the right now, much less trying to match wits with a Sicilian...I mean a computer with superior AI.

I am sure we all can agree that we are not ready to face Judgment Day just yet, well of the machine type anyways. So on behalf of mankind, womankind, humankind and beastkind-- I thank you for knowing when to say "when". (and thanks again for the great info)
Pirate Overlord
Mar 16, 2012
10631
wanted468 on Feb 5, 2015 wrote:
But Chok' Akab comes before you get Ratbeard. Chok is on the quest for Ratbeard's ghost crew.

Quentin Davenport, lvl 65

Marcus Davenport, lvl 25
Since the original poster is a swashbuckler ( as the player has Fan Flanders ) then protect her with Subodai.
Petty Officer
Feb 27, 2009
94
anecorbie on Feb 5, 2015 wrote:
Since the original poster is a swashbuckler ( as the player has Fan Flanders ) then protect her with Subodai.
I'm satisfied.
Quentin Davenport, lvl 65
Marcus Davenport, lvl 25