Exploring dynamic hazards to help balance games

Exploring dynamic hazards to help balance games

Postby tempAnon093 » 24 Feb 2019, 06:23

One issue that has been noticed in online games is that some more practiced and skilled players (who I will call 'expert' players) are able to complete tracks much more quickly than regular players.
There has been some discussion of how to help balance online games, which you can find in this thread. One of the main issues is figuring out how the game can be unfair to expert players without making them feel cheated or bored.

One idea I've proposed is using temporal or dynamic obstacles in the maps which will affect expert players more than other players. I think this could be a very useful technique, if well implemented. Discussing this among the STK community could result in interesting ideas to help liven up maps.

Basic examples of moving obstacles already exist, such as the large cannonball in Fort Magma and the train in Black Forest. These hazards can be coordinated so they occur more often earlier in the game or lap and occur less later. This means that an expert player who gets there early has more hazards to dodge that may slow them down, while a slower player has less to avoid.

One example is a place in the track with falling rocks, where more rocks fall earlier in the lap. Like the cannonball in Fort Magma, the rocks would bounce on the track before leaving so that later players won't be affected. The rocks could be replaced with something else depending on the track theme, maybe stalactites, golf balls or car traffic.

A more complicated example would be creating a shortcut that expert player are discouraged from using, that will allow less skilled players to catch up. I have drawn up a simple track map that demonstrates what this could look like.
Image
In this track, there is a shortcut that is completely blocked by a slow train for a while, forcing expert players who arrive early to take a more challenging route. Once the train has finally passed, players can now take a shorter route that will reduce the gap between them and the expert player/s.
I have also added a second train at the end of the shortcut. This means that an unsporting expert player who simply stops and waits for the train to move should not beat an expert player who immediately takes the long route. An expert player is disadvantaged by taking the shortcut, but slower players are rewarded, as shown by the small table in the image. If designed well, this kind of alt route could help force more balanced games even if the players have very different skill, without boring the expert player.

Challenges:
These are not easy, fool-proof solutions. Some potential issues I can think of are:
1) a hazard shouldn't discourage good players to drive well, but it should challenge them more. This may a hard balance to find. Why shouldn't they just stop and wait for the hazard to pass and then drive well?
2) if a track layout allows an expert player to overlap a slower amateur player, a very difficult hazard built to challenge the expert player every lap could affect the slow player! That would be especially mean and insulting.
3) timing and placement of hazards is critical so that expert players are penalized more than others. Badly-timed hazards would be worse than no hazards. Maybe triggers on the track could be used for timing when hazards begin and end.
4) An expert player could memorize a safe route through difficult hazards, or the timing of when a hazard stops. Randomization through scripting would prevent this to keep it challenging.
tempAnon093
 
Posts: 70
Joined: 02 Feb 2019, 12:09

Re: Exploring dynamic hazards to help balance games

Postby Auria » 25 Feb 2019, 00:02

Hi, this is not necessarily a bad idea, but this would most likely require a huge amount of work to implement. Considering that STK has few developers and even fewer 3D modellers or level designers, this is probably not realistic at this point
Image
User avatar
Auria
STK Moderator
 
Posts: 2972
Joined: 07 Dec 2009, 03:52

Re: Exploring dynamic hazards to help balance games

Postby tempAnon093 » 25 Feb 2019, 12:00

Auria {l Wrote}:Considering that STK has few developers and even fewer 3D modellers or level designers, this is probably not realistic at this point

Yes, for sure. I'm not making this post expecting people to spend time on it because I said so, I just think it's an interesting idea that might be useful in future tracks.
I am thinking of making a proof-of-concept track like this when I have some more spare time, but like my other track it won't be anywhere near the STK-assets standard.
All the parts needed are already used in existing tracks, I'll just have to adjust the animations and put them in place.
tempAnon093
 
Posts: 70
Joined: 02 Feb 2019, 12:09

Re: Exploring dynamic hazards to help balance games

Postby Alayan » 25 Feb 2019, 17:11

The focal point of the thread you linked to was that a player having a good start enjoys a lasting advantage, as being much less likely to be attacked by a bowling ball/cake/swatter means that the advantage often increase, despite the basket ball being there to somewhat regulate this (but with the bug which made it target other players + the ability of top players to dodge it for a long time...).

Here you're exploring something else entirely.

I can't help but think that the fundamental idea is wrong : the only way to give a bad player a shot at defeating a good player, you would need to make the game very random and unfair. Admittedly, losing repeatedly to a stronger player isn't the most fun of experiences, but :
1) The stronger players are also much more invested into the game. 24 players have done 100 or more ranked race. 15 of them are in the top 24 in rankings by point too.
2) In non-ranked servers, stronger players can and (often do) enable handicap, which helps make races closer. In ranked servers, there is a small incentive to use handicap but only if it doesn't compromise much winning (it also works fine if two top players use it at the same time so neither get an advantage in their direct match up).
3)When luck is only a minor factor, each and every upset is so much more satisfying.

In the online rankings, Gnu889 and Andet are around the 7400 points mark (and if I did more races, my own rating of 6800 would also converge to higher values). This means that theodorepringle, which is 4th in the rankings, is expected to win only about 25% of the time to them while losing 75% of the time against one of them. Mv81, which is 6th, is only expected to win about 13% of the time. As you can see, this is already a big skill gap and Mv81 usually loses. But when Mv81 manages to pull off a great race and to beat one of the top 3 players, it is very satisfying. He knows he earned the nice victory. If instead STK was made a luckfest ala MKWii, we could just brush it off "you were lucky this time" - there would be much less satisfaction - not to mention the rare but frustrating occurrence where he would outplay us only to lose because of bad luck. The same applies for other good players like slowcooked who can sometimes snatch a nice win against us.

A 7400 ranking also corresponds to a 98% expected likelihood of finishing ahead of a 4000-ranked players (which are already better than average !). That is, the difference in skill is so overwhelming that even when the 7400 player does poorly for his standards and the 4000 very well for his, it is still unlikely that the 7400 finishes behind. The amount of luck you have to introduce to give the less skilled player a 10+% likelihood of winning would make races between more closely matched players very random.

Items given to payers in worse positions, like zippers, are already a good help to catch up if used properly, but the "zipper crash" is way too common among unskilled players. As said in the other thread, I do think this item balance could use an improvement.

Now, how to fight the frustration of newbies losing too much online race?

Ultimately, the best solution would be for STK online races to be popular enough that it would be possible to match make players by skill : with some ranked servers reserved to players under some rating, and some others to player over another rating, we make sure that newbies wouldn't have to directly take on drivers way too strong for them. Many competitive online games use some sort of matchmaking system by skill to allow all players to have a shot at winning without having to compromise the core gameplay by making it excessively luck-dependent (this even work for games with no luck at all like chess).

These hazards can be coordinated so they occur more often earlier in the game or lap and occur less later. This means that an expert player who gets there early has more hazards to dodge that may slow them down, while a slower player has less to avoid.

You can coordinate them for the first lap, and that's about it ? You could make that triggering the lap line would trigger another set of dynamic hazards I guess.

Something which really should be avoided is making "doing nothing at the start" a remotely viable tactic. In MK, waiting for half a lap before catching up with items is quite doable...

Your example with the train is not a good idea at all in my opinion :
- what happen for a "middle level" player which take 25s to go to the point where the blocking train is ? Option 1, he goes to the expert path, but it will likely take more than 30s as he is likely to crash once. Option 2, he stops for five seconds to wait for the train to pass and then use the easy path. But while he will arrive at the end faster than if he had taken the expert path, all his effort to get to the fork in the road 5s earlier receive zero compensation, as players having taken 30 seconds lost no time compared to him.
- Player A and B are both top players, but player A had a much better 1st lap than B and is 10 seconds ahead. The path with the train is closed, but as the train is synced with player A, when player B arrives the easy path opens up. It takes player B 25 seconds to do the easy path, saving him 5 seconds and, with additional help of some zippers, he goes on to catch up and win. Is that a deserved win ? In my opinion, this is not fun. I had sometimes strings of defeat against Gnu889 and Andet when playing less well, but by tightening up my driving I could get a serie of wins. With tracks having many mechanics like this train path one, this would lose some of its meaning.

I'd love STK to have dynamic hazards, that is things which move and whose position affects optimal trajectory (making path memorization slightly less important), but the way suggested in this thread to help unskilled players to fare better is in my opinion not a good idea.
Alayan
STK Moderator
 
Posts: 197
Joined: 25 Mar 2012, 22:11

Re: Exploring dynamic hazards to help balance games

Postby tempAnon093 » 25 Feb 2019, 22:47

Alayan {l Wrote}:I can't help but think that the fundamental idea is wrong : the only way to give a bad player a shot at defeating a good player, you would need to make the game very random and unfair.

I personally don't think that being very random and unfair is the only way or even a good way to help balance the game. I think a small amount of those things can help compress the field (like the basketball can do) and add variance to a track, but they shouldn't be powerful enough to make the game more luck-based than skill-based. My thought behind the ideas I gave were to try and reduce the gaps between faster players (I've noticed in some of your videos, two or three players are completing the lap in half the time of other players!) and slower players. As you mentioned, a player who gets a good start and drives far in front of others "enjoys a lasting advantage, as being much less likely to be attacked by a bowling ball/cake/swatter", which just means the winner keeps winning. If the slower players are given a shortcut, for example, that will not put them ahead of the fast player but will close the gap from 8 seconds to 3, it means that most items can now be used against the lead players since you can actually see them, reducing the lead player's total domination from a good start. I think this is a good thing in a game mode with items.

Alayan {l Wrote}:Ultimately, the best solution would be for STK online races to be popular enough that it would be possible to match make players by skill : with some ranked servers reserved to players under some rating, and some others to player over another rating, we make sure that newbies wouldn't have to directly take on drivers way too strong for them.

Yes, I agree that would be nice. Meanwhile, I guess ranked matchmaking could be done manually with named servers. A description could say that maybe all players are welcome but this server has very experienced 4000+ players. Or is that too segregated for the amount of players we have?

Alayan {l Wrote}:Something which really should be avoided is making "doing nothing at the start" a remotely viable tactic. In MK, waiting for half a lap before catching up with items is quite doable...

Absolutely, which is what I tried to counter with the train idea, but as you pointed out there are important scenarios where it causes more harm than good. I did not think of it from a competitive point of view so I did not realize how easily it could make competitive players feel cheated by the unfair advantage given by the delayed shortcut.
I think a big part of the problem is that the train shortcut is a big, on-off shortcut. There is no consideration of middle cases like a medium-high level player and it is rewarding enough to make "do nothing at the start" (on purpose or by accident of a bad lap) the best route for a highly skilled player. It could be better on a far smaller scale, just like most STK shortcuts are, but it still is a binary solution to a non-binary situation, so it is not a great idea for balancing games.

Alayan {l Wrote}:I'd love STK to have dynamic hazards, that is things which move and whose position affects optimal trajectory (making path memorization slightly less important), but the way suggested in this thread to help unskilled players to fare better is in my opinion not a good idea.

Yes, I think that having more dynamic hazards in general would be nice. They don't have to be like the ones I described, and clearly the ones I have described are just plain undesirable to players who want to enjoy a mostly fair, skill-based match. I am curious to see if other players see value in the idea or similar ideas though, but yes it would not be a good idea to make a bunch of levels that are horrible for competitive players to race on.


As GhostX pointed out in another thread, some existing hazards like the golf balls (Minigof) and mine carts (Old Mine) are annoying as they are not advertised and more-or-less seem random to a driver. I think dynamic hazards like that which almost can't be dodged are also something that should be avoided. It just feels unfair. Hazards like the cannonball in Fort Magma would be great; it looks like it affects the optimal driveline but still gives players time to avoid it without it being useless. Are there any existing hazards you think are good examples?
tempAnon093
 
Posts: 70
Joined: 02 Feb 2019, 12:09

Re: Exploring dynamic hazards to help balance games

Postby Alayan » 01 Mar 2019, 04:39

tempAnon093 {l Wrote}:I think a small amount of those things can help compress the field (like the basketball can do) and add variance to a track, but they shouldn't be powerful enough to make the game more luck-based than skill-based.

I completely agree.

tempAnon093 {l Wrote}:My thought behind the ideas I gave were to try and reduce the gaps between faster players (I've noticed in some of your videos, two or three players are completing the lap in half the time of other players!) and slower players. As you mentioned, a player who gets a good start and drives far in front of others "enjoys a lasting advantage, as being much less likely to be attacked by a bowling ball/cake/swatter", which just means the winner keeps winning.


These two elements are (mostly) unrelated. If we take players who are closely matched in skill, for example Gnu889, Andet and me ; we'll complete laps in broadly similar times. But the one who manages to be first after 10 seconds will very often win the race. The first position lasting advantage doesn't require skill differences to manifest. Of course, when those differences exist, it adds up. But even when hit a lot by cakes/ball/swatter the good player will still be able to pull ahead of newbies before the end of the race. I think for

tempAnon093 {l Wrote}:If the slower players are given a shortcut, for example, that will not put them ahead of the fast player but will close the gap from 8 seconds to 3, it means that most items can now be used against the lead players since you can actually see them, reducing the lead player's total domination from a good start. I think this is a good thing in a game mode with items.

I focused more on the "gain 5 seconds for free" than on the "allows to attack the player ahead".

I prefer to give a zipper (or some new speed boost item) than a free shortcut for slower players.

My general ideas to limit first position advantage :
- the dynamic basket ball speed is intended to give a chance to dodge the ball for some time, but the dodging duration happens frequently to be too long with the best players. It is also weird when you can get two or three basket balls who can't catch up. So this may benefit from some tuning.
- In some tracks, the basket ball explodes against track elements, this should be fixed.
- Reduce the effectiveness of nitro denial. If you are about 2 seconds behind, it's too far to do proper attacks with bowling balls or a cake, but you can't take the nitro cans that the first player has, so even with better driving you're stuck behind.
- Reduce somewhat how much players behind attack each other.

tempAnon093 {l Wrote}:Yes, I agree that would be nice. Meanwhile, I guess ranked matchmaking could be done manually with named servers. A description could say that maybe all players are welcome but this server has very experienced 4000+ players. Or is that too segregated for the amount of players we have?

So far in the beta, you sometimes can get a 8 players server filled up but usually when you get players it's less than that... Hopefully we will get more online players with the main release, but realistically, we won't have enough for such a scheme to be effective. :| I hope long term this will be different. People have more fun playing against people of similar levels, it's not fun to crush newbies nor it is fun to be crushed by an expert player who has done hundreds of online races.

tempAnon093"It could be better on a far smaller scale, just like most STK shortcuts are, but it still is a binary solution to a non-binary situation, so it is not a great idea for balancing games.[/quote]
Indeed.

[quote="tempAnon093 {l Wrote}:
Yes, I think that having more dynamic hazards in general would be nice. They don't have to be like the ones I described, and clearly the ones I have described are just plain undesirable to players who want to enjoy a mostly fair, skill-based match. I am curious to see if other players see value in the idea or similar ideas though, but yes it would not be a good idea to make a bunch of levels that are horrible for competitive players to race on.

As GhostX pointed out in another thread, some existing hazards like the golf balls (Minigof) and mine carts (Old Mine) are annoying as they are not advertised and more-or-less seem random to a driver. I think dynamic hazards like that which almost can't be dodged are also something that should be avoided. It just feels unfair.


I agree !

The minigolf ball especially is something I have often discussed with Fouks and Andet as being very annoying and random.

tempAnon093 {l Wrote}: Are there any existing hazards you think are good examples?


In STK, I don't think we have good dynamic hazards anywhere. The fort magma ball is ok because it is avoidable, but it always fall at the same spot and thus it's not that dynamic.

But if we look at Mario Kart, we can get many ideas : enemies slowly moving in random direction, big dangerous elements moving according to a clear and predictable pattern, moving platforms following a pattern... :)

Maybe we can come up with something original, but adapting one of those dynamic hazards type into a STK track and with STK's theme would be a great first step I think.

Thanks for your post, it's good to have a lively discussion to find good ideas.
Alayan
STK Moderator
 
Posts: 197
Joined: 25 Mar 2012, 22:11

Re: Exploring dynamic hazards to help balance games

Postby XGhost » 02 Mar 2019, 17:11

Alayan {l Wrote}:Thanks for your post, it's good to have a lively discussion to find good ideas.

Gonna queue in here. Thanks to both of you. Really interesting and valuable for future track designers to read through this. It's always great to have some prolific opinions and results of the analysis from experienced players and developers here!
User avatar
XGhost
 
Posts: 132
Joined: 02 Jan 2014, 15:49
Location: Siebnen (Switzerland)

Re: Exploring dynamic hazards to help balance games

Postby tempAnon093 » 29 Apr 2019, 02:00

Apologies for the slight necrobump, but I just realised the Bovine Barnyard addon has the rolling haystack. I think this is a good example of a moving hazard which is slow enough to dodge but big enough to be a significant hazard.
tempAnon093
 
Posts: 70
Joined: 02 Feb 2019, 12:09

Who is online

Users browsing this forum: No registered users and 1 guest