• Topic Archived
You're browsing the GameFAQs Message Boards as a guest. Sign Up for free (or Log In if you already have an account) to be able to post messages, change how messages are displayed, and view media in posts.
  1. Boards
  2. Pokemon Card GB2
  3. Depressing news

User Info: MAtt5TER

MAtt5TER
11 years ago#1
We are know the AI in this game sucks. But, how much does it suck? I went and did an experiment.

I challenged all the Ghost Master, the supposed the best in the game, three times. These are the decktypes they use:
Wigglytuff Haymaker
Zam Swap
Trainer Restriction
Aggressive Fire
Balanced (Really, it's the best I can describe it. It has elements of stall, Haymaker, status, and power)
Legendary Zapdos (it's not Zapdos FTK)
Dark abuse
Sleep
Raindance

I set Dratini Lv10 (base set, 40HP) as the active Pokemon and vending Mr. Mime Lv20 (50HP) as the only benched Pokemon before my opponent's first turn. I would do nothing on my turns and count how many opponent's turns it took to knock out one Pokemon. Mr. Mime is there because its PokePower removes all weaknesses and resistances.

Results:

Note: Excuse the bad formating. The rows correspond to the list above so that the numbers in the first row (8, 2, 4) are the number of turns it took for my first Pokemon to be knocked out against the Wigglytuff Haymaker deck over the 3 trials.
Note: In all trials, Dratini was the one to be knocked out.

trials.......|averages | average deviation
1---2---3 |
___________________
8---2---4 | 4.67------ | 0.89
4---6---5 | 5.00------ | 0.33
x---3----x | n/a-------- | n/a
6---3---4 | 4.33------ |1.11
2---2---5 | 3.00------ |1.33
3---8---6 | 5.67------ |1.78
3---4---3 | 3.33------ | 0.44
2---3---3 | 2.67------ | 0.22
3---3---3 | 3.00------ | 0.00

x = I won; my opponent decked out -_-;

average number of turns it took to knock out the first Pokemon, not including the 3rd row = 3.96
mean average deviation of all trials, not including the 3rd row = .763889

Analysis to come on another post.
This sentence is as meaningless as you.

User Info: MAtt5TER

MAtt5TER
11 years ago#2
So, it takes 4 turns for the AI to knock out a Pokemon with 40HP. It averages out to be 10 damage per turn, which is pretty bad considering that you are supposed to powerup your active Pokemon as the turns go on. For comparison, Haymaker and Rain Dance decks could have easily KO'd that Dratini on turn 2. Only 19% of trials showed a 2 turn KO.

What's even more shocking is that the AI can't even win a game when I'm doing nothing. As seen in the 3rd row, the AI lost when I was trying to lose.
It makes you wonder whether the AI was really programmed to win or not.

Here's another number: mean average deviation = .763889. This number is supposed to show how consistent the in-game decks are. That is a pretty high number. While taking 4 turns to KO a Dratini is bad, the AI can't even do that consistently. Taking 5 turns to KO a 40HP Pokemon is a common sight. This is a deck issue. It shows that in-game decks are bad because they can't even follow a strategy all the time. For comparison, Haymaker and Rain Dance decks can consistently KO that Dratini. Their mean average deviation would be near 0.

So...thoughts? comments?
This sentence is as meaningless as you.

User Info: SSJ2__Gohan

SSJ2__Gohan
11 years ago#3
LOL yeah, the AI that uses the Trainer Restriction deck is a moron. Usually, he starts with Psyduck to Headache. Then he evolves to Dark Golduck and keeps wasting his energies on Third Eye unless he's down to 10 or fewer cards. Then, he gives energies to his Haunters and just lets Dark Golduck sit there O_o

Just wondering, did anyone use Kangaskhan or something else meant to merely set-up the bench (the Dark Golduck story excluded for obvious reasons)? Because in that case, it's not a fair comparison. Most of the Ghost Masters make a Pokemon a basic Pokemon not to attack you, but to set things up, and in that case it's unfair to call them "slow at dealing damage". The AI's terrible, but not THAT terrible.

User Info: MAtt5TER

MAtt5TER
11 years ago#4
Just wondering, did anyone use Kangaskhan or something else meant to merely set-up the bench (the Dark Golduck story excluded for obvious reasons)? Because in that case, it's not a fair comparison. Most of the Ghost Masters make a Pokemon a basic Pokemon not to attack you, but to set things up, and in that case it's unfair to call them "slow at dealing damage". The AI's terrible, but not THAT terrible.

Kangaskhan is a decent opener Pokemon because it has high HP and an ability to draw cards. It's a fairly good Pokemon to start with if you want to stall. I believe the AI is programmed so that high HP basic Pokemon have priority to be the opening active Pokemon. However, these Pokemon don't have priority in attaching Energy cards after the first Energy unless certain situations change the priority levels. Thus, Hitmonchan using only "Jab" and never being powered up is a common sight. However, this is a weakness of the AI. The AI can't adapt to different situations. It fails to notice that powering their active Pokemon is the better choice under the situation I presented. Thus, the AI is terrible.
This sentence is as meaningless as you.

User Info: SSJ2__Gohan

SSJ2__Gohan
11 years ago#5
Partly true. For example with Kanga, I don't waste my time giving him energies for Comet Punch, since before he can Comet Punch, his HP has usually dropped below an acceptable level (yep, against the AI), or is simply dead/poisoned/whatever. However, that's probably why you gave Hitmonchan as an example, right?
The partly true I was talking about, also goes for Hitmonchan (as for Kangaskhan, it's simply false). When using Hitmonchan as a starter, it all depends on how much he'll be weakened wether it's worth it going for Special Punch. After all, your benched Pokemon must be able to take over once Hitmonchan dies.
  1. Boards
  2. Pokemon Card GB2
  3. Depressing news
  • Topic Archived

GameFAQs Q&A