BLIND ALGORITHM + STATIC GEAR
Imagine a matchmaking algorithm that takes absolutely nothing into account.
It takes the first 10 players who queue up, splits randomly into 2 teams, and sends them off to battle each other.
If everyone had exactly the same gear (ships/weapons were all identical, and nothing could ever be leveled up)...
...then this is a system where skilled players can (and should) win more than 50% of the time.
... and poor skill players lose more than 50% of the time.
Poor skill players typically stop playing these types of games, and they go find another game.
(Not something most profit-seeking game companies want.)
BLIND ALGORITHM + GEAR PROGRESSION
Now imagine where ships/weapons aren't identical. Gear can be leveled up over time (i.e. progression).
If the algorithm is still "blind" to everything, then life for basically everyone gets even worse.
All the best gear may be concentrated on one team or another...
Skill begins to play a smaller and smaller role...
Gear progression dictates winning now.
Because if skill no longer determines the winner - it's just the LUCK from the blind algorithm for which team is stronger, and the stronger you are compared to the average, the more often that "luck" favors your team.
It's P2W heaven.
...and it's a recipe for failure.
...and game companies don't build games this way.
LADDER ALGORITHM + STATIC GEAR
Imagine having static gear again.
But now, as your higher skill causes you to win against average players, you will now climb a ranking ladder.
And then, the algorithm matches you against similarly-ranked players on that ladder.
Again, gear is static.
So, this is purely a skill-ladder.
Eventually, unless you are the #1 most skilled player ever, you will eventually rise (or fall) to a spot on that ladder where you no longer win (or lose) more than 50% of the time.
You will settle out at 50% win rate.
And all of this is amazing because all your games are against equally skilled opponents.
So, every game is fair, competitive, and fun!
LADDER ALGORITHM + GEAR PROGRESSION
And here we are. Our reality.
We already know from above, that if the game has Gear Progression, then a Blind algorithm will not be healthy.
So we need a Ladder-type algorithm.
The problem is, it won't be purely a skill-ladder anymore.
Now it's a ladder based on the combination of skill+gear.
But --- and this is very important ---
The algorithm needs to user their current place in the ladder (however they got there) to determine how they get matched.
Players will then navigate up/down the ladder naturally (based on how well they can utilize their own unique combination of skill+gear).
Let's call this a "natural ladder."
The algorithm should not try use pure gear to predict where they belong in the ladder, and then use that to change how they would otherwise have been matched.
If it did, that actually creates parallel ladders (each based on static gear score), where each match is like a blind algorithm for that gear level.
Which puts us back in the scenario where poor skill players lose a lot and quit.
So, should the algorithm use some gear to do a little bit of predicting?
No. It adds no value.
However... we still have problems in this game...
PROBLEMS
Even when we had a natural ladder algorithm (pure Infamy), strong players are often found to be lower in the ladder than where they "honestly should be".
Is it because their combination of skill+gear won't let them climb?
No.
It's because they are rewarded for something else.
Some incentives exist in this game that are directly tied to ladder rank.
But some Incentives exist in this game that are not directly tied to ladder rank.
If the non-ladder incentives outweigh the ladder incentives, then players find themselves improperly ranked.
When improperly ranked players have no incentive to climb, then in a natural ladder algorithm, players will be paired based on their current rank.
Which means we're essentially back to a game with GEAR PROGRESSION + BLIND ALGORITHM, where players who are vastly outmatched are still matched together.
We already showed that this is highly undesirable.
SOLUTIONS
So.... was the natural ladder the problem?
Or... were the imbalanced incentives the problem?
(Answer = the imbalanced incentives were the problem.)
Thus, the fix should NOT be to change the ladder to pure gear (obviously).
But the fix should also NOT be to change the ladder to a hybrid of position+gear either.
That would be trying to fix something that was never broken.
The fix SHOULD BE to correct the imbalanced incentives.
Ladder rewards should be vastly more desirable than non-ladder rewards (such as guild rivalry rewards).
Ladder rewards should scale up significantly enough so that players want to climb. No player should be able to say "well, if I drop 500 infamy the rewards are basically just as good."
If you do this, tanking (at any level) won't ever completely go away (it'd be like trying to convince everyone to never troll on Twitter ever again).
But it would drastically reduce.
Players will pursue the best incentives. Let's make those incentives encourage climbing.
You could even eliminate the awful rule where players who fall -800 Infamy stop earning ladder rewards.
Losing -800 Infamy is actually not outside the realm of natural possibility even when that player is trying their best!
It's an awful rule...
If they want to keep the rule, then make it a % of your highest infamy (like -50%), not a flat 800.
And make it so players below this threshold stop earning ALL rewards (ladder rewards AND guild rewards).
But I hate the rule altogether. It should be removed...
Once you fix incentives, and things improve, if there was any need to further prevent players from intentionally dropping (for troll reasons), then all you need to put in place is "gear score floors" where if you reach a certain gear score, you can't fall below a certain infamy. I've talked about these kinds of floors in other threads, and there's very little downside.
Ok, that's all I have time to write. Hopefully this has inspired some thoughtful consideration somewhere...
Click to expand...