Getting Started With The World Of Bats
The World Of Bats is a modding community and toolkit centered around bat simulation and habitat creation, primarily used as a side project within the broader wildlife simulation and indie gaming space. It lets you design custom bat roosting environments, model echolocation patterns, and run basic population simulations. The core toolset is free and open-source, which is why it has stuck around this long. You can grab it from the official GitHub repository. The main release page is at github.com/worldofbats/worldofbats. Grab the latest .zip for your OS. I use the Windows build myself. Don't worry about the Linux binary — it works, but the pathing on my machine was a pain until I stopped trying to run it through WSL and just used the native Windows version. Unzip it anywhere. There is no installer. Just open the folder and run BatSim.exe or worldofbats_gui.py if you want the Python-based interface. The GUI is slower but easier to navigate if you are new. The CLI version does everything the GUI does and runs noticeably faster.
I ran into a specific issue on my first install that took me about two hours to figure out. The simulation crashes immediately if your system locale uses a comma as the decimal separator instead of a period. The config files expect dot notation for all float values, and the parser just chokes. The workaround is to rename the config files in your user directory — they live in %APPDATA%\WorldOfBats\config.yaml — and manually swap every comma to a period. Alternatively, you can set your Windows region to English (United States) temporarily before launching. I prefer the config edit. It is less disruptive.
How The Simulation Actually Works
At its core, the simulator uses a agent-based model. Each bat is a discrete entity with properties like wing loading, echolocation call frequency, foraging range, and social nesting behavior. The environment is a grid-based terrain map where you place roost sites, water sources, insect prey zones, and obstacles. The simulation then runs tick-by-tick, and each bat makes decisions based on proximity, call return data, and energy reserves. One thing beginners consistently get wrong is the terrain resolution. The default setting is 10-meter grid cells, which sounds reasonable but is actually too coarse for accurate echolocation modeling. I dropped mine to 2-meter cells and the simulation quality improved dramatically, though runtime increased by about three times. If you are running a large map with 500+ agents at default resolution, expect it to take 40-60 minutes per hour of simulated time. At 2 meters, that jumps to roughly 3 hours per simulated hour on a mid-range machine. Another counter-intuitive detail: the echolocation physics in this tool are not real ray-tracing. They use a simplified cone-cast algorithm with a falloff curve. That means bats effectively "see" in wide conical sweeps, not pinpoint sonar. It is an intentional simplification for performance, but it means your simulation will overestimate detection accuracy at range. If you need precise acoustic modeling, this is not the tool for that. You would be better off looking at specialized bioacoustics software like Kaleidoscope or Sonoscan for actual field recordings. The World Of Bats is for behavioral and ecological modeling, not acoustic engineering.
Get the Full Details
Building Your First Scenario
Start with the included template called "Temperate Rooftop Colony." It has everything set up — a synthetic building structure, nearby insect spawning zones, and a basic parameter set. Open it, switch to the Parameters tab, and look at the foraging_radius and call_rate values. Those are the two knobs that matter most when you are trying to match real-world observations. I spent weeks once trying to make a simulation match actual field data from a brown bat colony in Pennsylvania. The numbers refused to converge no matter what I tweaked. The breakthrough came when I realized I was using call frequency defaults from a European dataset. The North American species use notably different frequencies, and the simulation's prey-detection math is frequency-dependent. Once I swapped the call parameters to match Myotis lucifugus specs, the model behavior aligned within 15% of observed patterns. That kind of specificity is easy to miss if you are just following the tutorial.
Exporting and Sharing
When you are satisfied with a run, you can export as a CSV, a JSON state file, or an MP4 visualization. The CSV export includes per-tick agent positions, energy levels, and call events. It is useful if you want to do secondary analysis in R or Python. The MP4 export renders a top-down view with trajectory trails. It looks clean but the file sizes are large — a 10-minute simulated run at 30fps comes out to roughly 800MB. The community share feature lets you upload scenarios to the built-in forum. It is low traffic but the people who post there tend to know what they are doing. I found a modified parameter preset for desert species there that saved me from having to reverse-engineer it myself.
Known Limitations
The simulation does not model disease spread, predation events, or seasonal hibernation cycles. If any of those are important to what you are studying, you will need to layer in external data or scripts. The developers have said hibernation is on the roadmap but it has not shipped yet as of my last check. There is also no multiplayer or collaborative editing. You are working alone unless you manually exchange scenario files. One more thing that trips people up: the tool assumes continuous insect prey availability unless you enable the seasonal depletion toggle. Leaving it off will make your bats never starve, which sounds convenient until you realize your population dynamics are completely unrealistic. Turn it on and calibrate the insect spawn rate to something that matches your target ecosystem. The default prey curves are guesses, not measurements.
