There needs to be an incentive to keep your rating up.

This is genius. Implement this. (As long as even fighting super low elo players/AI doesn't give bonus chests any more slowly than it does now of course).

 

That was an interesting article Spacedust, thanks for sharing it. Your were right to assume I wouldn't agree with it completely but I would be lying if I didn't admit there is a lot of merit in what he is saying.

I agree too that it's up to the devs to ensure that the game rewards fun play but I would argue that 'fun' is not actually that subjective and is very definable.

I applaud the honesty in people admitting to tanking, but in all fairness, there is no way to justify it. Why? Because if every high level player tanked then that means the chances of being stomped when your not ready increases. On the flip-side by quitting early you also deny the newer player the progression of climbing the ladder naturally. I think you severely underestimate the impact every one of these games has on rising players (granted you probably think I overestimate).

What is the source of the problem?

(imo) This is. High level play should be rewarding in itself. You should 'want' to play against higher level players because it's more fun to do so. If you dont, that is the problem. If we have a situation where players at the top tier feel like they cant compete without the very best equipment then the answer is right in front of us.

 

QFT. That's the natural incentive system that compels people to continue playing. That's one reason why I think draw decks are so despised - because it feels like a lot of 'unfun' work when you keep having to pass to let the other person continue his combo. I don't mind playing against them, personally, (especially if I end up stomping them) but I can see why many people dislike the drudgery of playing against them.

 

As long as you put any rewards/goals that is meta/outside of an actual match. The match will suffer. The goal of a single match must always be play to win no matter the costs. As long as there is a reward outside the match attached to it, the goal is slightly shifted away from "winning at all costs". It is guaranteed to happen.

If there is a reward for MP, there will be people playing for the reward, not playing to win.

If there isn't a reward for MP, nobody will be playing.

Let's see how another MP game solve this problem. Dota 2 rewards random item for player regardless of whether they win or lose as long as they don't leave the game. This make sure that players cannot determine any other goals (because its random) outside the match other than winning. Many other MP games also reward XP for completing a match that does not depends on win/lose (most of the time match duration).

In Dota 2 or SC2, the reward outside the match are purely cosmetic (though of great values) and does not affect the match. Card Hunter is not meant to be a serious competitive MP game, but its reward affect the actual outcome of future battles. This heavily weight the rewards much more than the outcome of current match for many players.

To solve the problem at the core, rewards must be:

- Non-deterministic: Random loot has already handled it, but it also need to be:
- Independent of win/lose status
- Cannot be gamed (quick resign etc...)

Comeback to the question of OP: No there is no need for an incentive for people to keep their ranking up. The current incentive for player to play the game should rather be reworked instead.

 

That's a very good point Progammer : The incentive to keep your rating down is what needs to be reworked, while having an incentive to play MP. Something like rewards depending on play time only (that would not encourage draging battles, as you would play less game in the same time)?

 

Right, because I have notice that some people will resign at the first sign of a disadvantage - it pays to make your games fast. If you lose, you will have a slightly easier win on the next game. With the rating system - in the end you will probably be 50/50 on wins and losses. So the incentive is to play fast and resign early right now to finish as many games as you can - about half should be wins. Making the reward time dependent would seem to solve many of the problems. I would think that people would try their best to win at that point.

edit: I see a thread about a 3 minute delay to get a new game if you resign in the first 2 rounds - that just might work

 

Why not ditch the rating system all together (as far as a matchmaker, at least) in favor of the SP level progression system? Have MP characters start at Lv1, and match players together based on level ranges of + or - 1 level, and if a high level player has no other opponents available in range, simply lower their levels enough to match with another player, just like SP when you engage a lower level adventure. That would easily prevent veteran players from stomping noobs, and would reduce efficacy in their decks when they do eventually duel a noob, as some of their card slots are removed from deleveling. For those who get bored with having all max level characters, have a, "Reset Level," button.

I never understood the reasoning behind MP starting at max level. It honestly kept me away because I had suspicions, that turned out to be true, that veteran players would have crazy decks that face-roll devastatingly well against new players. I play MP now, but I don't usually enjoy it, I just want to get some good loots and GP for my SP campaign. Due to my dislike of the way MP is setup, I have the strong feeling that once I complete the SP campaign, I'll simply walk away from Card Hunter for a few months, checking back occasionally to see if new SP content has been added, because MP is just way too overpowered.

Yea, I know, I need better cards and gear setups to compete well and rank up, it just worries me that I have to get super deep into the stratagem just to get some replayability. It seems there's no way for casual play with progress, as softcore players get stomped easily.

 

My suggestion would also require SP characters to be held separate from MP, or copies of them made strictly for MP, in order to work.

Side note: I can't figure out editing my messages, it takes me to a text window, but I can't type? I'm no forum noob, but this forum's design is a first for me.

 

That would not work at all : They'd still have epic/legendary focused low level items deck against ill fit starter deck , and would still wipe the floor with newbies. The current system works much better to match players against other players of similar skill and items.

 

I haven't seen every epic/legendary, but so far all the ones I've found from level 1-10 are not so good, or at least I've found commons with better cards. Keep in mind too that when you delevel for a fight, it not only removes slots, but removes tokens as well. I really don't think the current system works that well, because quite frankly 1/6th of my battles will be against a player who truly is on par with me, in terms of characters/items/skill, whereas the other 4/6ths of the time is against veterans who tanked their rating just to grind not-so-veteran players, and quite clearly outrank me in terms of items and general skill. 1/6th of the time for me the player either disconnects or resigns after round 1.

I've never had a good experience with any multiplayer game that employs matchmaking based on superficial ratings (as in ratings based on win/loss only). It would make more sense, to me, to have ratings doled out based on in-combat performance, such as number of kills, damage dealt, damage blocked, total turns to victory/loss.

No, those 1/6th of on par battles are not always won, just pretty equal in terms of strength of cards and general tactical ability. I would consider myself a darn efficient General in combat strategy games, I also play a lot of chess.

 

But it's not only based on win/lose : losing against a lower rated player drops your rating pretty badly, while winning against a higher rated one will give you a large boost. This is similar to the system used in chess after all. Elo is the only way to have more or less fair match, that is, once the incentive for tanking one's rating is removed.

 

Did you have any other examples of what the rating is based on? Even though you gain a boost from defeating a higher ranked opponent, and lose more from losing to a lower ranked opponent, it's still based on the fact that you either won or lost, it's just the increment that's based on the opponents rank.

Elo works well for determining a numerical rating, however you still have the issue of chests being awarded simply for a win, rather than how many rating points you took from the loser, thus the incentive for tanking one's own rating remains in place for the simple goal of quickly getting chests.

 

That's why I have been advocating handing chests depending on play time rather than win/lose condition.

 

The simplest solution I can think of that would keep the current rating system in place, is to change Resigns & Disconnects so that they result in no change of either player's rating. That way there, people can't simply tank themselves by quitting matches, and if an opponent is quite obviously trying to tank their rating by passing their turns, the other player can simply resign the match, which prevents the would-be tanker from achieving their goal. At that point, the only way one could purposefully tank their rating would be to get another player to agree to aiding them in tanking their own rating.

Of course, then you would have rage-quitters who would resign simply to spite the other player and prevent them from increasing their rating. A modified Elo system, where actions in combat grant extra rating bonuses, would solve the rage-quitting problem, to a degree, by allowing any bonus points accrued to be given out even with a resign or disconnect.

An alternative to granting Resigns & Disconnects zero change, is to result in no change for the player who resigns, while awarding points as usual to the opposing player who did not resign. That effectively removes tanking through resignation and disconnection altogether. I suppose then you could have the opposite problem, of a player agreeing to resign to boost the other's rating artificially, but then the player with the increased rating would be facing off against other players who are not tanking their rating, and likely would have much better items, making the artificer's games much more difficult and much less enjoyable. That's a self-correcting problem.

With either suggestion, the chests would have to be modified so that they are based on rating points earned, rather than a straight win, such as each chest requiring you to gain 5% of your total rating in order to be opened. 5% may be a little too steep, so maybe 2.75%, but then that might be too low, so try 3.25%
EX1 - 5%:
If your rating is 800, you need to gain 40 points for 1st chest, 840 requires 42 points for 2nd, 882 requires 44 points for 3rd (44.2 rounded to nearest whole).

EX2 - 2.75%:
If your rating is 800, you need to gain 22 points for 1st, 822 requires 23 points for 2nd(22.605 rounded to nearest whole), 845 requires 23 points for 3rd (23.2375 rounded to nearest whole)

Ex3 - 3.5%:
If your rating is 800, you need to gain 28 points for 1st, 828 requires 29 points for 2nd (28.98 rounded to nearest whole), 857 requires 30 points for 3rd (29.995 rounded to nearest whole)

As you can see from the calculations, the higher your rank, the more points you would need to gain a chest, although that would be a hindrance to high ranked players, it also curbs the disparity between low and high ranked players. Another alternative is a sliding % for chest requirements, with the % getting lower for each chest.

An example of a sliding % for chest requirement:
1st chest at 2.75% with rating of 800 requires 22 points, then 2nd chest at 2.7% with rating 822 requires 22 points (22.194 rounded to nearest whole), then 3rd chest at 2.65% with rating 844 requires 22 points (22.366 rounded to nearest whole)

Of course you will likely never get the exact amount needed to open the chest with the first battle, so just calculate the chest requirements after the previous chest was opened, and hold that amount until it is reached.
Ex:
1st chest at 2.75% with rating 800 requires 22 points. You gain 17 from first battle, 12 from second which surpasses the requirements for the 1st chest (29>22). Now at 829, calculate the 2nd chest at 2.7% for a requirement of 22 points (22.383 rounded to nearest whole).

Players could always tank their rating by playing through the entire match and losing on purpose, but in that scenario with my alternate suggestion (Resign & Disconnect yield no change for resigner, but points for non-resigner) in place, I would simply resign the match to spite the would-be-tanker, because at that point their rating just went up when they wanted it to go down, while my rating stayed the same.

My suggestions for changing the points doled out from Resigns & Disconnects may or may not be a good idea, but I really think I'm on to something with point requirements for chests and calculating them on a sliding scale. Whew, that was a mouthful.

Ideas?

 

Hmm, mulling over it some more I can see that players who lose a few matches will suddenly have a point requirement that is higher than they would have based on their rating. Perhaps if a player loses their match, the chest requirement is then recalculated with their current rating, so then losing matches is not punishing to getting the chest. You could adjust points gained towards the chest when you win a match and then lose a match by calculating the % of progress on a chest and applying that % to the recalculated chest.

Ex:
rating 800 chest requires 22 points, you win and gain 12 points, so the chest is at 12/22 points (54%) then you lose 3 matches (-36 points) and your rating is now 776 (800+12=812, 812-36=776) and the chest recalculates t0 21 points (21.34 rounded to nearest whole), apply the chest progress of 54% to the new chest requirement.
21*0.54=11 (11.34 rounded to nearest whole) so your recalculated chest is at 11/21 points (54%)

I tried to show how strongly it could be affected by losses, but now it seems to me that you would have to lose a TON of matches for it to really affect the numbers.

The sliding % may seem a little too complex, so perhaps you could just have a flat rate of 25 points per chest, and save the progress from each win with no detractions from losses. Meh, I tried. Simple is better.