[GUIDE] Following the Golden Rule

[Tiftid]

Lately I've been thinking of writing another guide, since it's been the traditional 8 months since my last one.
I have a kind of pipe-dream agenda of eventually writing a guide for every gamemode, even some unofficial ones. I think that would take around 7 years?

But I've already covered the most fun ones - A/D, Payload and KOTH. Why should I then have to waste 7 years writing negative-Nancy guides where I repeat myself over and over again saying "this gamemode is shit but you can make a fun map for it if you wrangle your layout in this specific way"?

Instead, I came up with a golden rule that would let me write a single guide to cover every gamemode all at once.
Get ready:

"A team should only be rewarded irreversible objective progress after winning no more or less than 2 teamfights in a row."​

Well, that should cover it. Thanks guys, make sure to tune in again in 8 months!
... In all seriousness, I realise that this might need some demystification.
So let's launch into the gamemodes!

---

A/D​

Initially, BLU controls the lobby (after a small inconsequential skirmish outside their spawn) and RED controls the forward hold.

1. BLU is attacking the forward hold.
   
   1. If successful, they take the forward hold area and have won 1/2 teamfights.
   2. If unsuccessful, they get pushed back to the lobby area and have won 0/2 teamfights.
2. BLU is using the forward hold area to attack the backward hold.
   
   1. If successful, they take the backward hold area and cap the point. Since they've won 2/2 teamfights, they're rewarded with irreversible objective progress - RED can't take the control point back from BLU.
   2. If unsuccessful, they get pushed back to the lobby area and have won 0/2 teamfights.

For RED, there isn't any irreversible objective progress, besides running down the timer and winning the round. However, running down the timer is something BLU can reverse if they do manage to cap the point before RED wins, because either time gets added or BLU wins the round instead.

I get the sense that some of you are not yet convinced that 2 teamfights in a row is necessary. So here's some insight from Valve themselves:

(SNIPPET FROM DEVELOPER COMMENTARY) [Jakob Jungels] Invulnerability also sharpens pacing by helping a team push through a defensive stalemate, when the enemy team turtles up and refuses to come out of their base. With a lack of defensive skirmishers, it's easier for the offense to build up the invulnerability charge without interruption.

The way I interpret this is basically:

• Uber acts as a way to force a team to win a teamfight. It's a countermeasure for A/D maps that give RED a single extremely strong backward hold.
• The existence of uber means RED is forced to push out of their backward hold and hold a little further forward.
• If RED has the forward hold, they're able to prevent the uber from physically moving far enough to let BLU take their backward hold.
• But, in the act of doing this, RED is forced back to the backward hold, giving BLU the chance to win the next teamfight and cap the point.

Here are some additional tidbits I've extrapolated from this:

• If the map makes it impossible for RED to have a forward hold, they can never stop a BLU uber from reaching their backward hold, so there's no goal for them to play towards and no matter how skilled they are, they'll lose once the BLU medic right-clicks which is no fun.
• If RED has two forward holds instead of one, it becomes pathetically easy for them to stop an uber from reaching their backward hold. So, RED should never be able to push into the lobby area and use it as a second forward hold.

This knowledge has basically informed the entire first half of my A/D guide, and is the reason why in my opinion exactly two teamfights in a row is an absolute necessity, at least in A/D.
Let's see if other gamemodes follow the same rules!

Not covering Payload, A/D CTF, VIP and A/D Pipeball because they function too similarly to A/D.​

---

KOTH​

KOTH is kind of unique. Neither team has irreversible objective progress besides winning the round.
Capping the point is essentially reversing enemy objective progress, since you're both stalling their timer and advancing your own, closing the gap between your team's scores.
That's why it's okay for the point to change hands even after only winning a single teamfight.

In fact, in KOTH you want to encourage the point to change hands as often as possible. You don't want to make it so that the team who wins the initial midfight will go on to win the round, because that would be irreversible objective progress after just one teamfight.

Here are some numbers that typically make for good KOTH maps:

• It should take 12 seconds to walk to the control point after spawning
• The control point should have a captime value of 6 seconds
• Teams should start the match with a respawn wave time of 6 seconds
• If your team owns the control point, your respawn wave time should be 8 seconds
• If the enemy team owns the control point, your respawn wave time should be 4 seconds

And a general map design guideline: you should aim to keep players as safe as possible up until they reach the capture point. So, a lot of cover, a lot of health, and relatively close-together routes so you're close to your teammates and aren't getting flanked by enemies.

Not covering King of the Flag since it (the good version anyway) functions too similarly to normal KOTH​

---

3CP (symmetrical a la Powerhouse)​

Initially, each team controls their own last point.

1. RED and BLU are both attacking mid. Let's say BLU are our protagonists here, cause if Payload pub pushes are any indication they definitely are.
   
   1. If successful, BLU takes mid and begins attacking RED last. They've won 1/2 teamfights.
   2. If unsuccessful, RED takes mid and BLU is pushed back and forced to defend their last. They've won 0/2 teamfights.
2. BLU owns mid and is attacking RED last.
   
   1. If successful, BLU takes RED last. Since they've won 2/2 teamfights, they're rewarded with winning the round.
   2. If unsuccessful, BLU gets pushed back to mid and is forced to defend it, essentially resetting them to the start of Step 2-2. They've won 0/2 teamfights.

This is all well and good so far. Even if BLU fails to capture RED's last point, they're able to rally back and capture it if they win the teamfight at mid, then win another teamfight at RED's base.
But let's flip the script so RED are our protagonists, and let's see where all my complaints with 3CP arise.

1. RED and BLU are both attacking mid.
   
   1. If successful, RED takes mid and begins attacking BLU last. But we know how this round plays out, and RED doesn't take mid.
   2. When unsuccessful, RED is pushed back and forced to defend their last. They've won 0/2 teamfights.
2. RED is defending their last point.
   
   1. If successful, RED pushes BLU back and begins attacking mid. They've won 1/2 teamfights.
   2. If unsuccessful, BLU takes RED last. But let's bend the fabric of reality so RED wins this teamfight instead (another thing I want to see happen in Payload pub pushes).
3. RED is attacking mid, which is owned by BLU.
   
   1. If successful, RED takes mid, pushes BLU back and begins attacking BLU last. They've won 2/2 teamfights.
   2. If unsuccessful, RED gets pushed back to their last point. They've won 0/2 teamfights.

Do you see the problem?
Even though RED has won two teamfights in a row, they haven't won the round yet.
In fact, they haven't done anything irreversible! If they lose the next teamfight, then lose the following fight at mid, BLU will take mid and then RED will be straight back in the dumpster, defending their own last point.
Just getting out of your own base and taking mid is the same amount of effort as winning the round from the initial game state where no one owns mid.
Which is the same amount of effort as taking an A/D point.

In fact, if you lose the initial midfight, you actually need to win three teamfights in a row to get irreversible objective progress.
One at your own last, one at mid and one at the enemy last.
This is too much effort, and if you've been in playtests with me on a 3CP map, you know I like to complain about this.
Losing the midfight means I'm essentially doomed to defend my own base until the timer runs out.
There's no win condition for our team once we lose the midfight, because we have to win three teamfights in a row which is plainly and simply too much effort.

So how do we fix 3CP?​

Well, here's a very interesting tidbit:
My map tc_babel is essentially an attempt to fix 3CP!
It does this via a few mechanics:

• The mid point is removed so that teams don't have to waste time capturing it before they can capture the enemy point
• Both teams start with a 10 second respawn wave time
• It takes you exactly 10 seconds to walk from your point to the enemy point (on every stage!), so you'll always be able to get there before they respawn and come back
• To prevent it from being too easy to win the round after winning the initial midfight, the enemy team gets their respawn time shortened to 4 seconds when you touch their point, and the point's captime is long (set at 11 seconds; with a good number of players you can start capping it, kill the freshly respawned enemy team as they try to defend their point and cap it before they respawn again)

These mechanics are... a little extreme, but they ensure that after you fight the enemy team off your point, you'll be able to go to their point without having to fight them at mid - as long as you make sure to actually kill them instead of letting them retreat.

Layout-wise, mid is made as disadvantageous to be in as possible - literally no health, relatively tight, plenty of cover for the people entering mid but not so much for the people inside it. This ensures that the enemy team won't perpetually stay alive in mid and pressure your base - you'll be able to kill them, and once you do you'll be able to push their point.

Some rounds still stalemate, but this is mainly because not enough teamfights happened where nearly everyone died. This is okay - not every teamfight will end in a bunch of deaths, but plenty do, and plenty of rounds end by capture because of this.
And, importantly, balanced teams will not always stalemate, because even if the teams have equal skill levels, one team can still easily fuck up and all die to a strong uber.

Not covering Tug of War, Territorial Control, PASS Time, Invade CTF or Pipeball since they all function very similarly to 3CP​

5CP is also very similar except it takes three teamfights to win from a neutral state and up to FIVE to win from behind, no wonder matches never end unless the teams are really stacked​

---

Capture the Flag​

Normal, Valve-style

1. RED and BLU roll out and meet at mid.
   1. If successful, RED pushes into BLU base. They've won 1/2 teamfights.
   2. If unsuccessful, RED gets pushed back into their base. They've won 0/2 teamfights.
2. BLU respawns and meets RED outside BLU spawn.
   1. If successful, RED pushes further into BLU base and grabs the flag. They've won 2/2 teamfights.
   2. If unsuccessful, RED gets pushed back into their base. They've won 0/2 teamfights.
3. BLU respawns again and meets the RED flag carrier outside BLU spawn.
   1. If successful, RED escorts the flag back into their base and captures. They've won 3/2 teamfights.
   2. If unsuccessful, the flag is dropped and BLU starts to defend it. RED has won 0/2 teamfights.
4. BLU is defending their flag from RED outside BLU spawn.
   1. If successful, RED escorts the flag back into their base and captures. They've won 1/2 teamfights.
   2. If unsuccessful, the flag is dropped and BLU starts to defend it. RED has won 0/2 teamfights.

All over the place. It takes three teamfights in a row to get irreversible objective progress if you're just starting from a neutral game state, but if you grab the flag and then lose a midfight, you can get irreversible progress after winning just one teamfight.

"Easy in, hard out / Chin CTF" (essentially a CTF map where mid has one-way routes into the enemy intel room, so you can grab the flag immediately after winning the fight at mid, but can't escort it out of the enemy base without going the long way past their spawn and facing another fight)

1. RED and BLU roll out and meet at mid.
   1. If successful, RED pushes into BLU base and grabs the flag. They've won 1/2 teamfights.
   2. If unsuccessful, BLU pushes into RED base and grabs the flag. RED has won 0/2 teamfights.
2. BLU respawns and meets the flag carrier outside BLU spawn.
   1. If successful, RED escorts the flag back into their base and captures. They've won 2/2 teamfights.
   2. If unsuccessful, the flag is dropped and BLU starts to defend it. RED has won 0/2 teamfights.
3. BLU is defending their flag from RED outside BLU spawn.
   1. If successful, RED escorts the flag back into their base and captures. They've won 1/2 teamfights.
   2. If unsuccessful, the flag is dropped and BLU starts to defend it. RED has won 0/2 teamfights.

This model is better in that it takes the golden 2 midfights in a row to win from a neutral game state, but it still has the same issue of "even if RED fails and drops the flag, they just need to win one teamfight to get irreversible objective progress" which just makes it unfair for the team whose flag gets grabbed.

How do we fix this?​

• Some maps employ "shot clock mode", where the flag's return time won't reset when it's picked up - this means that if RED tries to get the flag when it's outside BLU spawn and fails, the flag will return before they get to try again
• I personally prefer shortening the return time to 20 seconds, because this means the flag will simply return to BLU base if RED loses a fight while holding it - it can never get stuck outside BLU spawn and force them to defend it in an unfair scenario where one lost teamfight means they get capped on
• There's a mechanic called "touch return" which basically does the same thing (if BLU touches the dropped flag, it returns to their intel room), but again I prefer a 20 second return time cause it ensures that BLU can't return the flag without actually killing all the RED players trying to go for it

Not covering Mannpower since it's too similar to standard CTF​

---

Cloak-type (KOTH with no timer)​

Short captime:

1. RED and BLU are both attacking the control point.
   
   1. If successful, RED captures the control point and wins the round, even though they've only won 1/2 teamfights.
   2. If unsuccessful, BLU captures the control point and RED loses the round.

Long captime:

1. RED and BLU are both attacking the control point.
   1. If successful, RED gains cap progress on the control point. They've won 1/2 teamfights.
   2. If unsuccessful, BLU gains cap progress on the control point. RED has won 0/2 teamfights.
2. RED has cap progress on the control point. BLU respawns and attempts to clear RED off the point.
   1. If successful, RED finishes the capture. Since they've won 2/2 teamfights, they win the round.
   2. If unsuccessful, BLU takes the control point and begins regressing RED's capture progress. They've won 1/2 teamfights.
3. BLU regresses RED's cap progress back to nothing. RED respawns and attempts to clear BLU off the point.
   1. If successful, BLU gains cap progress on the control point. They've won 2/2 teamfights, but haven't won the round yet.
   2. If unsuccessful, RED gains cap progress on the control point. BLU has won 0/2 teamfights.

So, if you have a too-short captime, the game will end after one teamfight, and if you have a too-long captime, the team who loses the initial midfight will be at a huge disadvantage because they have to regress the enemy team's cap progress, similar to how they have to cap mid in 3CP.

How do we fix this?​

• Use some logic trickery to wipe the point's cap progress whenever a team leaves the cap area
• Dramatically shorten enemy respawn time when your team is capping the point, and have quick walk times to the point out of spawn, so the point's captime can be short-ish without it being too easy to rush captures

---

Player Destruction​

Good PD will essentially function exactly like KOTH. The longer you own the capzone, the better since you can keep depositing points, but if the enemy team captures the capzone, they can begin to reverse that advantage by depositing points while stopping your team from depositing any.
So, there's only really one rule for PD that is unique from KOTH. Don't make the map too wide and flanky, or it becomes too hard to attack the capzone and too hard to prevent your teammates from randomly dying and feeding the enemy team points.

---

Robot Destruction​

RD is essentially a fusion of PD and CTF. If you control mid, you can score points by destroying enemy robots, while preventing the enemy team from doing the same.
Since it's CTF, grabbing the enemy flag is typically too hard, requiring three won teamfights in a row.
So good RD map design will probably focus in hard on the KOTH part.
Although, if you do want to encourage the CTF part, you don't really need to be scared of making it too easy to grab the enemy flag, since even capturing the enemy flag isn't irreversible objective progress - it just gives you a lot of points.

---

Payload Race​

2CP (where you have to push your cart from your base into the enemy base):

1. RED and BLU push their carts to mid.
   
   1. If successful, RED stops BLU's cart at mid and pushes their cart into BLU's base. They've won 1/2 teamfights.
   2. If unsuccessful, BLU stops RED's cart at mid and RED is forced to defend their control point. RED has won 0/2 teamfights.
2. RED is pushing their cart towards BLU's control point.
   
   1. If successful, RED wins the round. They've won 2/2 teamfights.
   2. If unsuccessful, BLU stops RED's cart at their control point and pushes forward to mid. BLU has won 1/2 teamfights.
3. BLU and RED meet at mid.
   
   1. If successful, RED pushes into BLU's base and starts attacking BLU's control point. RED has won 1/2 teamfights.
   2. If unsuccessful, RED gets pushed back into their base and BLU starts attacking their control point. BLU has won 2/2 teamfights.

Again, much like 3CP. If the enemy ever pushes their cart into your base, you then need to win 3 teamfights in a row to win the round, which is just not feasible.
But you can't shorten the map, because that would make it too easy to win the round just by winning the initial midfight.

Side-to-Side (where each team has to push their cart from one point along the map's line of symmetry to another, a la Pipeline Stage 3):

1. RED and BLU roll out to mid.
   
   1. If successful, RED stops BLU's cart and begins pushing their own towards the control point. RED has won 1/2 teamfights.
   2. If unsuccessful, BLU stops RED's cart and begins pushing their own towards the control point. RED has won 0/2 teamfights.
2. RED and BLU roll out to mid. RED is pushing their cart towards the control point.
   
   1. If successful, RED pushes their cart to the control point and wins the round. RED has won 2/2 teamfights.
   2. If unsuccessful, BLU stops RED's cart and begins pushing their own towards the control point. RED has won 0/2 teamfights.
3. RED and BLU roll out to mid. Both carts are close to the control point.
   
   1. If successful, RED pushes their cart to the control point and wins the round. RED has won 1/2 teamfights.
   2. If successful, BLU pushes their cart to the control point and wins the round. BLU has won 1/2 teamfights.

Step 3 is the major problem. Since the carts can't roll back, they will eventually reach a state where they're both near the end of the track and a single won teamfight will win the round.

How do we fix this?​

• For 2CP PLR, it can help to use map logic to open doors to a shortcut route once the cart reaches the enemy base, so it's easier for the enemy team to go straight to the enemy control point after defending their own one.
• For Side-to-Side PLR, you can add a long rollback zone at the end of the track so the cart will roll back after a loss to a point where it would take two teamfights in a row to capture.
• Similarly, you can do a train finale like Banana Bay, where when your cart reaches the end of the track it calls a train and you have to defend your cart until the train arrives and blows it up.

---

Arena​

Arena is another special case. The objective is kiling the entire enemy team. This usually happens after just one teamfight, but in case it doesn't, the control point will unlock and you can capture that to win the round.
However, it's not like you can get ubered into and lose through no real fault of your own like on an A/D map with no forward hold, because in arena medics will usually not have uber until the second teamfight.
The only real way you can lose through no fault of your own is if your teammates all die one by one and leave you alone against a massive enemy team.

How do we fix this?​

• You should avoid making the map too wide and splitting the teams up too much. The worst arena maps are the ones where each team splits into halves on either side of the map, and if the other half of your team loses their fight you get encircled and killed no matter how strong your half is.
• Graveyard-style arena is a lot better than normal arena, because even if your teammates are terrible and die they can still be useful by healing you, and you get buffed if you're the last living player, so a single good player can legitimately win even if they have terrible teammates.
• I think arena maps should have health packs. The control point already stops any excessive camping, and having to bitch about who has to play medic/engineer is not fun. It's less of a priority if your map is Graveyard-style, but still a good thing to give the living players.

---

Special Delivery​

Special Delivery involves taking one neutral objective to another, so it's very similar to side-to-side PLR, although still different in some significant ways.

1. RED and BLU roll out to the australium area.
   
   1. If successful, RED grabs the australium and starts to take it to the rocket. They've won 1/2 teamfights.
   2. If unsuccessful, BLU grabs the australium and RED rolls out to the rocket. RED's won 0/2 teamfights.
2. RED and BLU roll out to the rocket area.
   
   1. If successful, RED captures the australium and wins the round. They've won 2/2 teamfights.
   2. If unsuccessful, the australium drops and BLU is forced to defend it. BLU has won 1/2 teamfights.
3. BLU is defending the dropped australium at the rocket.
   
   1. If successful, RED captures the australium and wins the round. They've won 1/2 teamfights.
   2. If unsuccessful, the australium resets and whichever team happens to be lucky enough to be there grabs it. BLU has won 2/2 teamfights.

The part where it falls apart is where the australium drops at the rocket. It takes way too long to return, so RED will respawn and come back before it returns, adding an extraneous teamfight that BLU has to win, which is unfair for them. Additionally, when it returns RED can easily just grab it again.

How can we fix this?​

• Make it so that the australium's return time is 10 seconds, then make it so that the australium is disabled for 30 seconds after returning - this means that not only does BLU no longer have to defend it at the rocket, but RED can't immediately grab it after it returns - BLU has a chance to rotate back to the australium spawn area and defeat RED and regrab the australium.
• Sadly, even if you do this, BLU still has to win three teamfights in a row after losing the first one, because they have to win a fight at the rocket, then at the australium spawn area, then at the rocket.
• This is why I personally prefer an australium with a 20 second return time, and after it gets picked up from its spawn point (after being returned, not just after being dropped) the control point locks for 60 seconds. Additionally, after the control point is captured, the australium is disabled for 60 seconds.
• This ensures that no matter what, a team always has to win two teamfights in a row to capture the australium - one to grab it, then one to hold it long enough for the control point to unlock.
• You may also want to include anti-flag barriers so that teams can't take the australium back into their own base and hold it with no chance of losing it until their team happens to win a fight at the control point.

---

3CP (symmetrical, a la Standin)​

If you're not familiar with this gamemode, it has three neutral control points along the map's line of symmetry, and owning all of them at any given time makes your team the victor.

1. RED and BLU roll out to the three points.
   1. If successful, RED captures 2/3 points while BLU captures the other one. RED has won 1/2 teamfights.
   2. If unsuccessful, RED captures 1/3 points while BLU captures the other two. RED has won 0/2 teamfights.
2. RED rolls out to the control point they don't own. BLU tries to capture one of RED's points.
   1. If successful, RED captures BLU's final control point and wins the round. RED has won 2/2 teamfights.
   2. If unsuccessful, BLU captures one of RED's two control points, although RED captures the control point they originally set out to. BLU has won 1/2 teamfights.

The major problem is that step 2-2, where BLU trades one control point with RED, can very easily repeat infinitely, since it leaves BLU in the exact same position they started in. So it creates a nightmare scenario similar to powerhouse 3CP, where if you lose the first midfight you're behind for the rest of the game and it's only a matter of time until the enemy team gets a good push and caps your last point.

So how do we fix this?​

• Well, you could try using tf_logic_cp_timer to lock the two points that RED captures, then when the timer expires you give one of them back to BLU for free, so immediately after defending their final point, BLU gets to attack RED's final point. I tried this in a microcontest map and it worked... well-ish? despite the layout I made for it being terrible.

TCDOM (as seen on Caper) is similar enough to normal DOMCP that I don't feel like it's worth its own section.​

---

 

[Sonoma]

 

[SnickerPuffs]

Jokes aside, this is actually a really useful guide that explained something I didn't realize I noticed.

 

[Waffe]

I agree on the concept of forward and backward holds, but how would you define a teamfight? I would argue, since you don't count the push right at the start of a round as a teamfight, that badwater's first point is a single teamfight point (take control of the highground), and it works just fine.

 

[Tiftid]

It can definitely be hard. I remember back during the development of sd_nowhere I had trouble deciding where one teamfight ends and the next one begins.
Initially, I was relying on very strict time periods, but eventually I decided to relax it, and decide that the first teamfight only concluded once the flag was picked up, and the second teamfight only concluded once the flag was returned or captured.

I think for Badwater it mainly comes down to this factor: On Dustbowl, the small skirmish outside spawn has a lot more significance than just a small skirmish, because RED's forward hold really is right outside BLU spawn, and BLU uses their spawn as their lobby - quite clever, since RED physically can't push into BLU spawn.
But on Gorge A, RED and BLU both spawn a lot further from the point, so it becomes necessary for BLU to use an extra area outside their spawn as the lobby.
This area is obviously not holdable/retakeable by RED. It wouldn't be fair if it was, and besides it's way too big and RED incurs ridiculous rotate times trying to push back into it and RED spawns way too far away to hold it, compared to how close BLU spawns to it. So it's a classic lobby area.

The way I define Badwater A is that the area juuuuuust outside BLU spawn is the lobby, the top of the hill is the forward hold and the bottom of the hill is the backward hold.

 

[Custard1]

Top tier guide, as per usual. I feel like I never took on enough of this for my later stuff. Canaveral is too far gone for ANOTHER rework but I have some later microcontest stuff I'll definitely consider this for.

 

[Waffe]

It can definitely be hard. I remember back during the development of sd_nowhere I had trouble deciding where one teamfight ends and the next one begins.
Initially, I was relying on very strict time periods, but eventually I decided to relax it, and decide that the first teamfight only concluded once the flag was picked up, and the second teamfight only concluded once the flag was returned or captured.

[...]

I think we all have our own definition of a teamfight which is why I think it is important to write down what the guide means by it. Would a good definition be that a teamfight is whenever a hold is being contested where, at the moment, it is not clear who controls the held area?

 

[Tiftid]

Yeah, that works as a definition. The important thing about teamfights is mainly that winning means you gain control of the contested area.

 

[Brokkhouse]

This is a great write-up and I'm surprised the two-teamfight-rule doesn't already have a more central spot in the Tf2maps knowledge canon. I'll certainly revisit some of my maps in new light.