The Future of Gems of War on PS4/Xbox

Just tried it and look like a great game but the loading time is crazy

re: Real PvP: I recently ran into this article about making live multi-player, and was looking for an opportunity to ask the dev ghods how accurate it was. Adding multiplayer to a huge video game takes longer than a week - Polygon also, LOL @ “AI is so stupid”!

The difficulty with discussing features is that most people simply don’t have the language necessary to discuss it properly. It’s not simply a matter of hard/easy possible/impossible - it’s all about what the codebase was designed to support.

But even ignoring the technical parts, let’s think about the player experience. Real-time PvP means you need to find a player with an appropriate team who is looking for a PvP match at the same time you are. You could easily be waiting a few minutes for the match to start, if you’re playing at a time or a level that is uncommon.

Right now matches are all client-side, which means you don’t actually ping the server until the match is over. That’s 1 network call at the end of the match. The servers right now are balanced for that level of load. Sending a network call to and from (2 calls!) after EVERY match/spell would be something like a 30x increase in network traffic per game! And that also means a 30x increase in the number of games dropped or lost due to poor connections.

And what if your opponent simply stops making moves? Or if you need to deal with something for a moment? How long does the game wait? Is there a new timer implemented? How does that affect your ranking score? Are disconnects taken into account during matchmaking?

There is also the potential problem of boards getting out of sync. That’s not possible now, since there is only 1 board. But with multiplayer - it could happen. So there needs to be some sort of provision to check that - which means even more network traffic. And code to resolve things if the boards are mismatched, etc etc etc.

13 Likes

I agree game will take much longer to complete, meen you get less gold and glory.
And i don’t want to wait 10 sec after every move, i don’t see anything good in real pvp match

2 Likes

Fascinating insight here. Before this, I thought, “Ehh, it could be cool some day, especially on console.” Now, it really doesn’t seem worth it. I can only imagine the costs for that much extra server bandwidth would be substantial.

3 Likes

nuzzles @Alpheon

The other thing to note is that this would likely exacerbate the stall-y situation we have going.

Imagine vsing Manticore/Khorvash/Mab - except played by a human. And someone would come up with something even worse, that has days’ worth of stalling with no progressive effects, the sole purpose of which is to troll people until they quit. Only it wouldn’t just be that one guy, it would be dozens of people.

Not that it would be any fun for most people to use that team, I’m just saying it’d happen eventually.


Turn-based games take much longer with real-time PVP than, well, actual real-time PVP. Compare Age of Empires II to Civ V or something. Even accommodating how horrifically slow I am at Civ V, it must still take at least four times as long to finish that compared to AoE. … I know they’re different games, just roll with it.

3 Likes

The closest thing to multiplayer experience is to have a heroes vs heroes mini-game. No monsters involved. And if that could work, we’re looking at a guild war mini-game to come next. Just a question of how to regulate that fairly and determine what rewards could be for the victor.

A better AI wouldn’t be too hard, but it would be a lot of work. Currently all the troops use the same AI decision processes. You would need a new AI decision for each troop, and maybe some based on traits.
5 match
4 match
Use a troop power
Skull
3 my mana
3 opponent mana
Those are about what we have now in order of preference but with a bit of randomness thrown in.

Take a fully charged valkyrie. You’d add a decision choice would I make a 4 or 5 match if I used power? Then make it more likely to cast. Or alternatively are there few of the colors I transform make it less likely.

Shegarra, if the board has lots of red and skulls use the power. If very few then wait.

A troop with Diamond skin trait. Maybe make it less likely to want to skull match.

Not hard to do, just lots to do lol

They do exactly that @UKresistance. There’s a base AI and another “list of priorities” with the Troop’s spell

Are you sure? Because Mr creates a single skull rarely hits.

You talking about Aziris?

The AI doesn’t know how to “set up” moves. Aziris’ spell is setting up a move in the programming perspective

3 Likes

I dunno if I’d call make a skull match a setup.

Exactly. The most annoying trait of the AI is passing up using a spell in favor of a 3-skull match. I’ve won many matches I would have lost if they had just fired off a spell. It would have killed the troop that ended up tipping the match in my favor.

2 Likes

It is when you consider you have to program an AI step by step.

When I use Azaris, I try to set up a 4 match skulls. When the AI uses him, there’s usually one skull on the board and it just throws its spell first chance it gets. I’ve seen Skeleton do a 3 skull match instead of throwing its “add a bunch of skulls, then get another turn”. the AI isn’t as “smart” on consoles as it is on PC/Mobile. (and yes, there’s reasons for those quotes).

3 Likes

The ability to not plan ahead vastly hampers the AI. I have dozens of scenarios where I have made the AI do what I want because I know it’s priorities and restrictions.

1 Like

[quote=“evilbatwitch, post:165, topic:11123, full:true”]
When I use Azaris, I try to set up a 4 match skulls. [/quote]

I use Aziris is a Purple-heavy, Skull-based team and just use him 4-5 turns in a row.

1 Like

Me too… (10 chars)

1 Like

Na, programming the ai like that would be easy. Most of the hard stuff is already done as the computer is always aware of all legal moves. Looking for a 2 skull match and creating a skull to make a 3 isn’t much different than identifying a legal move.
The added ai for each troop easy, but doing it for 200 troops time consuming lol