My Review - Why I Only Rate 3/5

toopa
Hello World
Hello World
Posts: 5

My Review - Why I Only Rate 3/5

Post#1 » 20 Jul 2017, 11:17

Hello, I previously planned to make a long review of this game with my device (Android) on the app store, but after an hour of writing/thinking all my text vanished, so I thought: "AHHHHHHHHH-HHHHHHHHHH *cries in a corner*" -> using a computer is a better choice

As you see, I only rated this game 3 stars of the maximum 5. My main point of this post is to reason that rating and to tell you what I'd like to see in the game (improvements, new ideas).
Note: I'm completely new to the forums and didn't read any posts. What I mean with that is that my point of view hasn't been influenced by (the opinion of) others and I may suggest things you are already aware of.


Basically, the game focuses on two things
    1) making your own AIs
    2) competing against enemy AIs (online)

My problem about 1)
First when I started this game, I couldn't put it away. It was so addictive to constantly tweak and improve your AI, to own your opponents. Now, I rarely play it anymore, and don't enjoy making new AIs at all. Recently I spent hours and hours to make new AIs or to improve them. The "code" I used made sense in theory (at least I think so), but in praxis it was quite the opposite. This means, that once you reached higher leagues and have to tweak your AI in the best possible way, it's not about finding the best rules for your AI by using logic, it's 99% about doing research and battle tests about every case. With that I mean, you need to find out how the game handles your bots in the most efficient way, if your rules make sense or not is irrelevant. To find out how to "trick" the game/enemy AIs most efficiently, you need, as already mentioned, do hundreds of tests and spent hours in research, but not making new AIs - which is the game's concept all along. If this wasn't frustrating enough, I play this game on my (relatively old) smartphone and it's laggy and slow to change AIs in testing - I'll mention this point later again.

My problem about 2)
This isn't your fault, but the player base is so low (on Android), especially on higher leagues. I constantly play against the same 4-5 guys. This really lowers my joy for this game.

These two points are the main reasons why I only give you a mediocre rating as of now. I'd like to rate 5 stars, as the concept of this game amazing and I always wanted a game like that, but I was a bit disappointed in the end. I think you can't do much about it, but I also have a few suggestions.


Suggestions

1) Map Editor - Custom Scenarios (in-game)
It would be amazing if we were able to make our own maps and scenarios. The player who publishes a scenario uses his own AI and the community tries to defeat it. You could also add stats to every scenario:
a) WR of the enemy AI (only counts the very first battle you do against it)
b) average score of people who defeated the AI with the first try
c) average score of people who lost against the AI with the first try

d) average percentage of people who managed to defeat the AI at least once
e) average number of tries people needed to defeat the AI for the first time
f) average score of people who managed to defeat the AI at least once

Additionally, it would be cool if you could sort scenarios regarding to the stats (highest WR, most tries needed to defeat at least once, etc). You then could look at the scenarios we made and (possibly) implement the ones you like best to campaign or practise. I think that this suggestion is very important to keep the game fresh and to give the community the possibility to make an infinite amount of content.


2) Small Improvements
Assigning a different AI to my bots causes the game to lag and it's EXTREMELY annoying to always manually change the AI of EACH bot. There needs to be a "assign AI to whole team"-button immediately. What's even more annoying is that the AI/"formation" you use when doing battle tests (in "test") is saved and used in campagin. While this can be practical for people who always end testing with their finished AI, it's pain in the arse for people like me who do a lot of (small) tests.


3) Practise Is Boring
Practise is really outdated and the AI is way too easy in most cases. Actually, you fight with shotgun, machine gun and sniper class against assault only. I wish everything was a bit more challenging and there were more campaigns accessible. In case you enable us to make and publish scenarios, you don't need to add more practise levels in my opinion


4) New Rules/Conditions
I'm not sure what your future plans are about new rules and conditions, but I struggle with a few things.
a) how to put the following information in one condition block? "if myself (health is 0-25% OR 25-50%) AND (shield is 0-25%) ... [add whatever conditions you need] ... exists". I think you can't and that's really annoying.
b) how to say "if closest enemy (is carrying a resource) AND (is close to enemy base) exists"?
c) how to say "if closest enemy (is at close range OR is at medium range) AND NOT (enemy allies in close range OR medium range) exists"?
d) how to say "if myself (is attacked by X enemies) exists"? - where X is an integer
I bet there are many more cases that are either awkward to implement or not possible at all


I think this pretty much is it, I hope I didn't forget anything from yesterday's text that I failed to save on the app store. If I can think of more suggestions, I'll update this post.
Greetings, toopa

Ein Stein
Automaton
Automaton
Posts: 131

Re: My Review - Why I Only Rate 3/5

Post#2 » 20 Jul 2017, 11:44

True, the player pool is small compared to other online games. We should think of a new idea to attract new players.

And next time, to prevent the words from vanishing, try typing in Word or Google Docs or other similar apps first, and then copy-paste in onto where you want to post.
We cannot solve our problems with the same thinking we used when we created them.

Arcueid_57
Script
Script
Posts: 41

Re: My Review - Why I Only Rate 3/5

Post#3 » 20 Jul 2017, 12:46

"it's not about finding the best rules for your AI by using logic, it's 99% about doing research and battle tests about every case."

Based on my experience, I have to disagree on that point,
I had kind of similar impression at a certain moment and I stopped playing for some months. When I restarted, it was from the scratch, with a new account and with no records of my previous AIs.
Then I took the time, even before writing a single AI, to define my priorities, per bot class, purely theoritically. Then I transcribe that into a generic AI (except for SYP) and applied it to all maps, just based on strategic theory and memories 3 months old. With that new AI, I reach after some days the Master League.
I'm sure if today I erase all my AI and restart again from zero, it will take time but it will benefit to my ELO score.

User avatar
LuBeNo
Autonomous Entity
Autonomous Entity
Posts: 532

Re: My Review - Why I Only Rate 3/5

Post#4 » 20 Jul 2017, 14:42

toopa wrote:1) Map Editor - Custom Scenarios (in-game)
It would be amazing if we were able to make our own maps and scenarios. The player who publishes a scenario uses his own AI and the community tries to defeat it.

I really like this idea.

toopa wrote:4a) how to put the following information in one condition block? "if myself (health is 0-25% OR 25-50%) AND (shield is 0-25%) ... [add whatever conditions you need] ... exists". I think you can't and that's really annoying.

I also think this should be one node. Currently only possible with many nodes.
Image
My algorithm of life: if(self.tired) sleep(); else if(self.hungry) eat(); else follow(Jesus);

pier4r
Skynet
Skynet
Posts: 3390

Re: My Review - Why I Only Rate 3/5

Post#5 » 20 Jul 2017, 14:49

"it's not about finding the best rules for your AI by using logic, it's 99% about doing research and battle tests about every case."

Disagree. In every competitive game you optimize in such way.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

TheKidPunisher
Autonomous Entity
Autonomous Entity
Posts: 565

Re: My Review - Why I Only Rate 3/5

Post#6 » 20 Jul 2017, 15:16

Only the 4)b) is impossible actually.

mcompany
Autonomous Entity
Autonomous Entity
Posts: 872

Re: My Review - Why I Only Rate 3/5

Post#7 » 20 Jul 2017, 17:09

First off, hello and welcome to the forums. Anyways
toopa wrote:Note: I'm completely new to the forums and didn't read any posts. What I mean with that is that my point of view hasn't been influenced by (the opinion of) others and I may suggest things you are already aware of.

Or just, you know, request things that has already been requested (not that I have anything against those who do, it is just that most requests at this point has already been said several different ways in several different place (even on the Google play reviews and my friends, both of with who don't keep up the game and know nothing about the community))
toopa wrote:This means, that once you reached higher leagues and have to tweak your AI in the best possible way, it's not about finding the best rules for your AI by using logic, it's 99% about doing research and battle tests about every case. With that I mean, you need to find out how the game handles your bots in the most efficient way, if your rules make sense or not is irrelevant. To find out how to "trick" the game/enemy AIs most efficiently, you need, as already mentioned, do hundreds of tests and spent hours in research, but not making new AIs - which is the game's concept all along.

Like almost everyone else, I can't quite agree with you here. First off, one of the main things this game promotes (although it's actually not quite the strongest tactic) is making only one AI to conquer everything, so to say that making a more efficient AI is usually more important than making a new AI, then that's how it should be. Second, naturally in any competitive game will end with more players investing more time into finding the best solutions to the game, and there is no other way to the top with the research into the game. Lastly, while you say making your own AI don't work, I'd like to disagree, because usually if one set of rules don't work, you could also update the rules with the AI. However, at a certain point, in order to maintain the AI you are working on, you must have some sort of rules you go off of. One example of people's rules for the AIs would be here: viewtopic.php?f=3&t=823
toopa wrote:1) Map Editor - Custom Scenarios (in-game)
It would be amazing if we were able to make our own maps and scenarios. The player who publishes a scenario uses his own AI and the community tries to defeat it. You could also add stats to every scenario:
a) WR of the enemy AI (only counts the very first battle you do against it)
b) average score of people who defeated the AI with the first try
c) average score of people who lost against the AI with the first try

d) average percentage of people who managed to defeat the AI at least once
e) average number of tries people needed to defeat the AI for the first time
f) average score of people who managed to defeat the AI at least once

Additionally, it would be cool if you could sort scenarios regarding to the stats (highest WR, most tries needed to defeat at least once, etc). You then could look at the scenarios we made and (possibly) implement the ones you like best to campaign or practise. I think that this suggestion is very important to keep the game fresh and to give the community the possibility to make an infinite amount of content.

+1
toopa wrote:2) Small Improvements
Assigning a different AI to my bots causes the game to lag and it's EXTREMELY annoying to always manually change the AI of EACH bot. There needs to be a "assign AI to whole team"-button immediately. What's even more annoying is that the AI/"formation" you use when doing battle tests (in "test") is saved and used in campagin. While this can be practical for people who always end testing with their finished AI, it's pain in the arse for people like me who do a lot of (small) tests.

viewtopic.php?f=2&t=831#p9558
toopa wrote:a) how to put the following information in one condition block? "if myself (health is 0-25% OR 25-50%) AND (shield is 0-25%) ... [add whatever conditions you need] ... exists". I think you can't and that's really annoying.

I'm surprised that no one has requested this, but definitely +1
toopa wrote:how to say "if closest enemy (is carrying a resource) AND (is close to enemy base) exists"?

It is not currently possible. What it would require is some way to know both what the enemy is doing and the exact target the enemy is doing that to, which not possible. This would be what could fix it, but the problem is designing what the node should look like to make it: https://trello.com/c/n9UycGQE/74-ai-tar ... efactoring
toopa wrote:how to say "if closest enemy (is at close range OR is at medium range) AND NOT (enemy allies in close range OR medium range) exists"?

See 4b. You can get around it (somewhat) though. If you are fleeing from, attacking, or moving to any enemy, you can check if enemies you are not doing that to exist
toopa wrote:how to say "if myself (is attacked by X enemies) exists"? - where X is an integer

You could use tags... but that can usually be slow to set up, and take large subtrees. Otherwise https://trello.com/c/9EmPkjQ4/18-condit ... ing-allies

TheKidPunisher
Autonomous Entity
Autonomous Entity
Posts: 565

Re: My Review - Why I Only Rate 3/5

Post#8 » 20 Jul 2017, 18:30

Also the player on PC and on Android are mix in the same ladder

toopa
Hello World
Hello World
Posts: 5

Re: My Review - Why I Only Rate 3/5

Post#9 » 20 Jul 2017, 20:47

Thanks for your replies

toopa wrote:What I mean with that is that my point of view hasn't been influenced by (the opinion of) others and I may suggest things you are already aware of.

mcompany wrote:Or just, you know, request things that has already been requested


That's basically what I meant as well, I just didn't point it out

mcomapny wrote:First off, one of the main things this game promotes (although it's actually not quite the strongest tactic) is making only one AI to conquer everything, so to say that making a more efficient AI is usually more important than making a new AI, then that's how it should be.

I probably used the wrong word. I tend to copy my AIs to have a backup and then modify them - also giving them a new name. After I added/changed 10-20 more nodes, I make another backup and give it a new name once again. That's probably why I used the wrong term here. Basically, my problem is that I'm losing the joy of modifying my AI because it can happen that you add like 20 nodes and it doesn't change much or makes things worse. Example: I want to make different, specific retreat patterns for all the classes. It's a lot of work to find out what the best "retreat values" are, especially when you take every situation into account (how many enemies attack me, how much health does the enemy/all enemies have, are allies nearby etc). I literally get overwhelmed by all the possibilites that I should implement to make my AI as good and flexible as possible, but that's so much work and testing, and I'm just not satisfied with a 50% working AI. And if you finally managed to do all that stuff, you realise it would have been better to sacrifice your bot in this situation, to do something different in the other situation. To me it feels like I'm always doing something wrong and this is kinda frustrating and not entertaining to me - it feels more like work (maybe my ego to be one of the best players is too high).

Maybe that's how the game is intended to work, but I say it like this: I started the game 10 days ago and now my peak is close to 1.700 ELO. At the very begginning, let's say I edited my AI for 1 hour. After I'm done, I go to campaign, fight and unlock next league in 10 mins (exaggerating a bit here). Then I face new players with way better AIs, and I have to do the AI-editing process again. This repeated in the past days multiple times and I spent a lot of time creating/improving my AIs, but actually barely spent time playing against players. Now, things got really harder and I struggle to get better. I realised that especially in the higher leagues, it's a lot of minor details that make a difference. How much time do you think it would need just to find those 2-3 tiny details that will turn utter defeats into convincing victories (it often seems to be very small things that make the difference)?

Then I started to do more fights, as I got pissed of constantly improving my AI. What happened? I played against the same guys over again and again. I even started to remember where I lost and won on most maps against the guys I played. I could regularly tell if I was to lose a fight on map X against player Y or not, because I already was in the very same situation multiple times. Let's improve the AI? Nah, before I do the very detailed changes, I'd rather spend a lot of time doing fights against MANY different AIs, not just the 4-5 people I always get matched up with or play on a new game mode, so I get more information.

I'm not sure if I'm just a whiny bitch but if I made a log I bet 85% of the time I played this game was editing my AI, 10% was doing fights and watching the outcome and 5% was watching stats in the stats tab, on which maps is my AI doing well,where it's bad, etc.

toopa wrote:it's not about finding the best rules for your AI by using logic, it's 99% about doing research and battle tests about every case.

pier4r wrote:Disagree. In every competitive game you optimize in such way.

First off, I want to let you guys know that I was a bit exaggerating there with the percentage, of course it's not that extreme. I fully understand your reply and it makes sense, but for me, the competition and optimization was there from the very first second, and increased more and more every time I reached the next league. I probably focused too much on making a great AI on the very beginning, trying to get a good WR etc. I think my biggest problem is my ego and my urge to have good stats, as I am a good player in literally every game I play (not trying to brag).


If there were different game modes or whatever (I'm fully aware that it takes a lot of time to implement them) then this would have never happened in the same way, as I would have had multiple challenges, stayed in my league longer and even if not, I could have tried different challenges, maps, scenarios... whatever. I assume it's basically my fault for being too focussed on having a good WR and defeating most players I face, but as of now, it feels like I was driving a Ferrari with high speed, passed a lot of other drivers and then crashed into concrete and ended up in a wheelchair.

TheKidPunisher
Autonomous Entity
Autonomous Entity
Posts: 565

Re: My Review - Why I Only Rate 3/5

Post#10 » 20 Jul 2017, 21:18

If you understand well the game reach higher league but, like in every game when you reach opponent of your level the WR will goes to average 50%.

Yes the game is more played for modify the AI than fight. To give you an idea, i was stuck at 1400 after few weeks of play. I stop do match and work on my AI during 1 month and it bring me at 1700.

Other thing, i dont know your AI (i dont remember to fight you) but i think that at some level, feneral AI doesnt work very well.

If you want to goes up, more specific map AI will be needed.

toopa
Hello World
Hello World
Posts: 5

Re: My Review - Why I Only Rate 3/5

Post#11 » 20 Jul 2017, 21:38

Yeah, you are probably right. That's another thing that bothered me a bit, that you technically have to make multiple AIs to have the best results. It's possible to make a good overall AI, but there are certain maps where you will have an easier time with an individual AI. My in-game name is ToFu96, but I usually use "toopa" in other games/forums

TheKidPunisher
Autonomous Entity
Autonomous Entity
Posts: 565

Re: My Review - Why I Only Rate 3/5

Post#12 » 20 Jul 2017, 21:50

So we have met 1 time on circle of death, not enough to understand your AI... and not enough for you to see my particularity.

If ypu stay in master maybe we will match more often.
(You also give me an idea that i will think about before put it in request forum)

Miojo
Script
Script
Posts: 37

Re: My Review - Why I Only Rate 3/5

Post#13 » 21 Jul 2017, 03:30

toopa wrote:Yeah, you are probably right. That's another thing that bothered me a bit, that you technically have to make multiple AIs to have the best results. It's possible to make a good overall AI, but there are certain maps where you will have an easier time with an individual AI. My in-game name is ToFu96, but I usually use "toopa" in other games/forums


I use a single AI and I'm doing great with it :)...

Also I have to disagree with you saying that it is not about logic/creativity... each AI is very unique on its behavior, and a lot is because of the goals of its creator. If it was just about tweaking mechanics for each situation AIs would behave very similar, and the best AIs would be the one with more nodes..... ( nice reference viewtopic.php?f=3&t=802 )...
I think you stopped on a point that seems that it is just mechanics ... my jump from ~1700 to ~2000 was basically mechanics.. but after you get to 2k+/1900+ and you get losses that you CAN'T pinpoint where your mechanics went wrong, then you must start thinking "how could i change the macro behavior of my bots so this situation does not occur".... sometimes it is in fact about changing 1 or two nodes. But 1 or 2 nodes that completely change de dynamic of the game.

Also it is really unfair to the dev to complain about problems that exists because of the relative low pool. I would love to fight hundreds of different opponents... giving the game 3 stars is not going to help achieve this goal.

I agree that the game needs better conditions, but I also think it needs to be added with a lot of testing.. if we had inifinity chain of conditions then the problems of "tuning for each situation" that you mention would probably rise.

3/5 for a game that you spent a lot of time and had good moments, that has NO pay2win element, only requires "skill", and has a single dev...... you must have very high standards....

Ein Stein
Automaton
Automaton
Posts: 131

Re: My Review - Why I Only Rate 3/5

Post#14 » 21 Jul 2017, 04:11

Miojo wrote:I use a single AI and I'm doing great with it :)...

You mean one single AI for every map, whether it's Best Score or Elimination?
We cannot solve our problems with the same thinking we used when we created them.

Miojo
Script
Script
Posts: 37

Re: My Review - Why I Only Rate 3/5

Post#15 » 21 Jul 2017, 04:14

Ein Stein wrote:
Miojo wrote:I use a single AI and I'm doing great with it :)...

You mean one single AI for every map, whether it's Best Score or Elimination?


One generic for all maps. No decorators for specific maps or score/elimination.

I have ~20 nodes that will proc 95% of the time only in mind games (nodes that will activate when there is no friend sniper around).

mcompany
Autonomous Entity
Autonomous Entity
Posts: 872

Re: My Review - Why I Only Rate 3/5

Post#16 » 21 Jul 2017, 05:32

I am the same (only the same AI for all maps) except occasionally with Mind Game

toopa
Hello World
Hello World
Posts: 5

Re: My Review - Why I Only Rate 3/5

Post#17 » 21 Jul 2017, 12:55

Miojo wrote:Also it is really unfair to the dev to complain about problems that exists because of the relative low pool. I would love to fight hundreds of different opponents...

It may seem unfair, but it definitely lowers fun factor and I have a point. If, let's say, CS:GO only had 10 players, would you give a good rating on Steam?

Miojo wrote:giving the game 3 stars is not going to help achieve this goal.

You are probably right with that one, I'll change to 5*

Miojo wrote:3/5 for a game that you spent a lot of time and had good moments

It's just a temporary rating (especially as this game is still in alpha) and I didn't spend that much time on the game yet, just saying. If this game had more players, it would be so enjoyable.

Miojo wrote:that has NO pay2win element

If it had any p2w elements, I would instantly rate 1*, maybe 2*, as I do with any other p2w game.

Miojo wrote:and has a single dev

I would say that literally nobody cares if a game was made by 1 or 100 developers and how much time they needed to create it. In the end, what really matters is the product and if the customers are happy with it.

And with that I'm not saying that I don't appreciate the developer's work and I'm fully aware that it is much work to make such a game and that it takes more time for a single developer to implement new features than for a team of 30 people.

Miojo wrote:...... you must have very high standards....

Probably yes, but if I stop enjoying a game that I just bought like 10 days before, that lacks in content, I don't see a reason why I should rate 5*.

Miojo wrote:you get losses that you CAN'T pinpoint where your mechanics went wrong, then you must start thinking "how could i change the macro behavior of my bots so this situation does not occur"

This sounds interesting (and frustrating) indeed. I'll give the game another try after vacation.


I have a final question. How was it with the player base in the past? I noticed a couple of bad reviews on Android were about "paywall" after the tutorials, were there more players in the past, as this game has 500.000+ downloads on Android?

User avatar
Kanishka
Skynet
Skynet
Posts: 1421
Contact:

Re: My Review - Why I Only Rate 3/5

Post#18 » 21 Jul 2017, 14:45

toopa wrote:I would say that literally nobody cares if a game was made by 1 or 100 developers and how much time they needed to create it. In the end, what really matters is the product and if the customers are happy with it.


This. Fact.
Fixes break an AI more than bugs do. :ugeek:

Gladiabots Off-Topic Chat


My Stats: Kanishka_RN3;Kanishka_MiPad

User avatar
Kanishka
Skynet
Skynet
Posts: 1421
Contact:

Re: My Review - Why I Only Rate 3/5

Post#19 » 21 Jul 2017, 14:48

toopa wrote:I have a final question. How was it with the player base in the past? I noticed a couple of bad reviews on Android were about "paywall" after the tutorials, were there more players in the past, as this game has 500.000+ downloads on Android?


No. Player Base is MORE than what it used to be. (Thanks to Google showcasing the game, and Dr D making to the Indie Game Contest Finale)... Paywall means that you actually got 10 multiplayer Matches and 5 practice Matches for free. After that you'd have to pay.
Fixes break an AI more than bugs do. :ugeek:

Gladiabots Off-Topic Chat


My Stats: Kanishka_RN3;Kanishka_MiPad

Ein Stein
Automaton
Automaton
Posts: 131

Re: My Review - Why I Only Rate 3/5

Post#20 » 21 Jul 2017, 15:05

Kanishka wrote:
toopa wrote:I have a final question. How was it with the player base in the past? I noticed a couple of bad reviews on Android were about "paywall" after the tutorials, were there more players in the past, as this game has 500.000+ downloads on Android?


No. Player Base is MORE than what it used to be. (Thanks to Google showcasing the game, and Dr D making to the Indie Game Contest Finale)... Paywall means that you actually got 10 multiplayer Matches and 5 practice Matches for free. After that you'd have to pay.

Change the paywall to 10 multiplayer matches per day and then I can recommend this game to my friends.
We cannot solve our problems with the same thinking we used when we created them.

Return to “General Discussions”

Who is online

Users browsing this forum: No registered users and 20 guests