I'm remaking Hydro as an MvM map, and I need help/feedback

  • Site Migration: See bugs? Report them here. Want something changed or have an idea? Suggest it here.
  • If you're asking a question make sure to set the thread type to be a question!
  • Something not downloading? Download authors read this.


(*single chuckle*)
Apr 10, 2014
An idea came to me a couple days ago-- Convert Hydro into a Mannhatten-styled Mann vs. Machine map.

It stuck with me, and I really want to do it, but it's admittedly a terrible idea. MvM and TC come from entirely different eras of TF2's development, and have different goals and philosophies for their level design. You can't just swap out the logic entities (which somebody already did) and expect a perfect map, you have to put thought into it.

The best way to do this, I feel, is to recreate the map from near-scratch. That'll make it easier to adjust the layout into a more MvM-bespoke state.

Given the notoriously massive size of the map, I'd rather not do this alone. Plus, I've only made one MvM map before this, and it was a group effort. So, I'm asking for help from the community, both with the brainstorming and with the actually creation of the map.

Here's what I have so far...

Each team's final points will be cut, and the map will feature a linear, horseshoe-shaped path for the bots to follow. RED spawn (and, by extension, the bomb hatch) will be at the Radar Dish. The first robot spawn will be at Dam, with gates at Power Plant and Warehouse that move the robots closer to the hatch. Defending players will have robot-free shortcut routes to-and-from each main area, which will close up as the robot spawns move forward.

Taking inspiration from Doppler, I want the capture progress to be controlled via map/mission logic (so if a mission-maker wanted robots to start at Power Plant or Warehouse instead of all the way back at Dam, they can set it up that way).


There's still plenty of other things I/we need to puzzle out, however, such as...
  • What areas/routes will need the most trimming down during the recreation/redesigning process? Which one will need to be opened up?
    • Futhermore, how should each area/route be redesigned to better fit the conventions of MvM's level design?
  • Should this map include tanks? Where would they spawn? Would they be able to capture gates, like in Dockyard, or should they avoid the gates altogether?
  • Most community-made MvM maps feature forward upgrade stations. Should this map have those too? Where would they go?
  • Should this map be compatible with the Potato's/SigMod custom features right-from-the-start?
  • This should go without saying, but what would the missions look like?
  • Does this map need a new name, since "mvm_hydro" is already taken? What should it be?
Last edited:


(*single chuckle*)
Apr 10, 2014
Just started on this a couple days ago, but I only put a few hours of work into it so far.

I began at A gate (Power Plant/entrance to BLU Last) and I'm working my way backwards to the inital robot spawn at Dam.

(The first screenshot uses Hydro's light/skybox, while the second and third use Void's sky_twilight_02 skybox)



The Asylum

The Clown Prince of Mapping
Apr 5, 2008
And you're building this from the ground up? Damn.

I think it looks good so far. I can't remember the last time I played MVM, but what you've got so far makes sense


(*single chuckle*)
Apr 10, 2014
I have a day-off tomorrow, so hopefully I'll do another big burst of progress like I did over the weekend.

Still, with my limited time these past few days, I've finished the main room of Power Plant, plus a bit of the area outside.


I was going to start on Dam next, but it just dawned on me that it's comprised of about 85% displacements. My displacement skills are very subpar, so instead I'll start working on Warehouse.


(*single chuckle*)
Apr 10, 2014
Warehouse is just about finished! Sure, all of the doors are static props, and the cliff walls are simple brushes because I'm a coward, but it's basically finished nevertheless!


Here's an overview of everything I have built so far, along with how far along I think the map is, percentage-wise
  • DAM - INITAL ROBOT SPAWN - 5% (Has basic "floor geometry" for the robot spawns)
  • POWER PLANT - FIRST GATE - 80% (The most complete area. Unfinished or missing connectors between itself and other areas)
  • WAREHOUSE - SECOND GATE - 70% (The cave route to Warehouse has been roughly covered up, and will likely need revisions. Also missing connectors)
  • RADAR DISH - RED SPAWN AND HATCH - 5% (The titular dish, copy-pasted from the SDK vmf of Hydro)
  • GAMEMODE LOGIC - <5% (Only robot spawnpoints, loosely following the naming conventions of Mannhatten's spawn points) (main0, main1, upper02, ect.)
  • OVERALL - 35% (Should be self-explanatory)


This is about the point where I feel most comfortable asking for help. Granted, I've been asking for help with this from the get-go, but now that I have a decent percentage of the basic geometry done it feels like an actual project. It was just an idea before.

If you're interested in helping, please reach out to me! You can drop a reply in this thread, shoot me a DM either here or on Discord, or show up to my house in real life.


(*single chuckle*)
Apr 10, 2014
I've been picking up more shifts at work (plus I recently got another of my classic "Half Baked Mapping Ideas" and I've been trying to puzzle out if-and-how I'm going to get it done) so my chances to work on this have been getting fewer and farther between.

I've started to chip away at RED's spawn, though! I don't have any interior lighting yet, so here's a Hammer++ screenshot so you can see what's going on in there.


A few days ago, it occurred to me that while I've been constantly asking for help with this, I haven't actually specified *what* I need help with. Ha ha, whoops.

So, after all this time, here's what I actually need help with, in descending order:
  • Geometry/Layout (Revising Hydro's layout to improve MvM-compatability and general flow, getting the map finished)
  • Logic/Navmesh (The gatebot logic system, and the Doppler-style mission-controllable gates)
  • Detailing (Moreso later-on than now, but still necessary)
  • Popfiles (@Tiftid and @zythe_ have already offered to create popfiles, and I plan on making one-or-two myself, but the more the merrier!)


✖ ✖ ✖
Feb 20, 2018
i can help with detailing, and i will do the pop file. i dont know what makes a good mvm map so i dont know if i can improve the layout in any way, and logic i dont know either. but i am confident in my detailing skills, though that would have to wait until after smissmas is over because im stressed out working on smissmas maps for now. also, small question, since the map will be so long are you planning to have engineer robot support?


(*single chuckle*)
Apr 10, 2014
but i am confident in my detailing skills, though that would have to wait until after smissmas is over because im stressed out working on smissmas maps for now
That's perfect, actually! I was planning on getting the first playtestable version of Rehydrate out by mid/late November, then spending a few weeks playtesting and iterating before starting a proper detail pass. If we're lucky, everything should line up nicely.

also, small question, since the map will be so long are you planning to have engineer robot support?

Yep! I even have a few spots in mind for where their nests will be set up.


the Embodiment of Scarlet Devil
Sep 10, 2016
I'm happy to do any geometry or detailing gruntwork you need. I don't remember agreeing to write popfiles for this map, but I'm always happy to write an MVM popfile for anyone in need of one, so I have no doubt that I did agree, and I have no intention of refusing now.

But what I mainly have to offer is my thoughts, so here you go:

How I feel about the idea:​

Converting Hydro to an MVM map is interesting, and will probably go over well in players' eyes, since Hydro is cool and people like to see it given a second chance at success. The idea of a 4-arena circular map with defender shortcuts and gates is also interesting - it doesn't feel like a 4-arena map to defend (no ridiculous walk times for defenders), but it maintains the ease of defending that a 4-arena map would have.
I think you would need to have very, very long waves to give the bots any chance of victory, so probably 4 or 6 waves instead of the 7 or 8 most people seem to prefer.

How I think it would play:​

ARENA1 (Dam / Bot Spawn Area): Bots take after 1 strong subwave
ARENA2 (Power Plant / Gate 1): Bots take after 1 strong subwave if they already control ARENA1, gamestate soft resets when bots capture this due to gate capture properties, so defenders can push forward and keep holding this arena
ARENA3 (Warehouse / Gate 2): Bots take after 1 strong subwave if they already control ARENA2, gamestate soft resets when bots capture this
ARENA4 (Radar Dish / Hatch): Bots take after 1 strong subwave if they already control ARENA3

So, each wave where the bots could conceivably win would need to have at least 6 strong subwaves (first captures ARENA1, second captures ARENA2, third takes ARENA2, fourth captures ARENA3, fifth takes ARENA3, sixth captures ARENA4). Using this example from Dayal's steam guide, Bavarian Botbash only has a single wave with 6 subwaves, and I doubt they're all strong subwaves.

That's basically why I recommended long waves.
But, it also acts as food for thought with the map's design. Hydro's areas are all reasonably easy to hold, but if you intentionally made it so that ARENA1 and ARENA3 aren't holdable (e.g. rotating between routes takes way too long so bots will usually get past the defence), you could make it so that defenders will:
  1. Initially hold ARENA2 and lose it to one strong subwave
  2. Continue holding ARENA2 even after the gate is captured thanks to the gate stun, then lose it to one strong subwave
  3. When RED loses ARENA2, they should usually lose ARENA3 because they have to respawn and they can't try to hold ARENA3 for fear of the bomb slipping past them
  4. This means RED will have to hold ARENA4 and lose it to one strong subwave
  5. If you think that's too punishing, you could make ARENA3 not be a gate so RED can still try to shore up their defence by holding ARENA2 even if they get pushed all the way back to ARENA4
With this new structure, you could make a more "normal" MVM map/mission where three strong subwaves are capable of pulling out a victory for the bots, and if one subwave captures ARENA2, the bots have more chance of getting through with the parts of the wave that only have two strong subwaves in a row.

But I don't really prefer it, since only having one gate isn't conventional, and having incredibly long waves with some tough subwaves would be something really distinguishing and unique.
For instance, I don't even think a 9-subwave wave would be hard on this map as you originally proposed it, as long as it was broken up by some weak subwaves which give the defenders time to push forward before the tough robots spawn.

Here's how I basically think it'd go:

Subwave 1 (HARD): Takes ARENA1
Subwave 2 (HARD): Takes ARENA2, bots get forward spawn at ARENA2
Subwave 3 (HARD): Takes ARENA2
Subwave 4 (EASY): Defeated by defenders, defenders push forward and retake ARENA2
Subwave 5 (HARD): Takes ARENA2
Subwave 6 (HARD): Takes ARENA3, bots get forward spawn at ARENA3
Subwave 7 (HARD): Takes ARENA3
Subwave 8 (EASY): Defeated by defenders, defenders push forward and retake ARENA3
Subwave 9 (HARD): Takes ARENA3

Now, one more HARD subwave right after Subwave 9, and the defenders' geese would be cooked. Similarly, if the defenders ever failed to push forward after defeating an EASY subwave, they'd have a hard time fighting off the following HARD subwaves. But on the flipside, if the defenders occasionally managed to defeat a HARD subwave with minimal deaths, they'd temporarily keep hold of whatever area they're in, and only push forward if they aren't defeated by the next subwave. So it kinda of keeps the normal balance?

How I feel about the layout:​

I think most of Hydro's arenas are good for MVM as-is. The connectors are worrying with their limited room to dodge robot attacks, but some of them are going to be one-way for defenders, and the other ones are only going to be used by defenders who are retreating anyway, so I think it's fine.

Usually, when modifying an official map, I recommend as few layout changes as possible, to avoid confusion from people who are familiar with the map. But Hydro is a special case, because very few people are familiar with it, and even the people who are familiar with the map tend to not really have a great mental picture of the connectors. So I think it's probably best to make the defender shortcuts as simple as possible - a relatively straight shot between the hatch and the gates, with an A and B sign above each one. For optimisation, you can make the tunnels dip down and then back up and use areaportals, so their straightness doesn't create a huge sightline that renders the entire gate area from the hatch area.

The shortcuts also introduce an interesting dynamic. The ARENA1 -> ARENA2 dynamic is just like a normal 2-arena MVM map (think Decoy) but a little harder for defenders - usually, on a 2-arena map, defenders only have to walk through 1 arena to get to the hatch, compared to the 2 they have to walk through to get to the bot spawn zone. But, with these shortcuts, defenders would have to walk through 1 arena (ARENA4) to get to the shortcuts, then one arena to get to the arena they desire. So it's easier for the bots to capture ARENA2 because of this.

I think to make it even more interesting you could remove the ARENA4 -> ARENA1 shortcut. That way, defenders have a 3-arena walk time to get to ARENA1, and a 2-arena walk time to get to ARENA2. But the bots only need to capture 2 arenas, so the probability that the first gate will be captured is high, and the game is less likely to just stagnate at ARENA1 like on Mannhattan, because if a defender ever dies, chances are they're not getting back to ARENA1 before the next subwave of bots has control of it. So defenders will usually end up actually having to retreat after losing a fight at ARENA1.

Removing that shortcut would also make navigation a lot simpler, because out of spawn defenders only have two options - go to capture point A, or go to capture point B.

Tanks and other minor stuff:​

I feel like tanks wouldn't be a particularly worthwhile inclusion. They're a mechanic that usually force defenders to retreat - they artificially weaken the defense in the area they spawn in by forcing 2-3 defenders to focus on them, making it easier for the bots they spawn with to kill the other defenders.

But in this map there are already gates which will naturally move the defense back over time. Tanks just feel like something that would add more confusion on top of a map already jam-packed with decisions - for instance, after you spawn, do you go to A (ARENA2) to help your team in the fight, or do you go to B (ARENA3) in case they lose and you really need to stop the gatebots/bomb carrier? Now think about if there was a tank in play too. It rapidly becomes mindless chaos.

Potato/Sigmod stuff: Nice to have if you can set it up, but I personally probably wouldn't use it in my mission(s).

Forward upgrade stations: Yes. Plonk them down right in the middle of each arena and don't turn them off mid-wave. If I want to upgrade mid-wave, I don't want to also have to yell at my engineer to buy 2ways, or worse, yell at my teammates to switch to engineer.

Overarching theme for missions: Idk, probably has to do with pyrosharks and fire because haha funny water get it?? Vague industrial theming would also work, like engineer bots (battle or otherwise) and demomen themed as "pipelayers" or "metalworkers" or something


(*single chuckle*)
Apr 10, 2014
I haven't read your entire post yet, since I'm getting ready for work. I'll go over it properly once I clock out for the day.

Although, given my intentions to have the gates controllable via mission logic (So, like Doppler, waves can start with the A or B gates pre-captured), the waves won't *always* need to be as long as you suggest.

I don't remember agreeing to write popfiles for this map, but I'm always happy to write an MVM popfile for anyone in need of one, so I have no doubt that I did agree, and I have no intention of refusing now.


Eh, it was less of an agreement and more of an off-the-cuff offer. It'd be awesome if you still made a mission, though!
Last edited:


(*single chuckle*)
Apr 10, 2014
  • I recently purchased a new and more powerful computer (because my old one died)
  • The recent TF2 updates have refreshed my interest in the game
  • One of the specific recent updates added an experimental feature that upped the maximum player count on a server to 100, not just for normal matches, but for MvM as well
  • Someone on Knockout! Forums reminded me about this, to which I responded "oh shit yeah i forgot about that"
All of these events have culminated in me wanting to get return to this gargantuan project. However, my initial "poking-through-the-VMF-to-remember-what-I-need-to-do" inspection threw up some roadblocks.

Firstly, a had converted several blocks of geometry into instances. My instances folder was lost when my old PC died. I didn't lose much, only a handful of buildings, but those are still sections I'm gonna have to remake entirely from scratch. Secondly, WOW a lot of this geometry is rougher than it needs to be, even for a "get it working now and make it good later" type of project.

Also, after I nudged Tiftid and Zythe last year, Tiftid reached out and offered to make the Dam area! He spent a few days working on it, but he was disappointed with his work so things quickly fell apart.

Since I'm renewing my commitment to work on this, I'd also like to renew my call for assistance. If you or someone you know is free and wants to pitch in, please reach out to me! The "done-ness" percentages I posted last october are still accurate!