WebTug Classic guide. The controls, scores, and shared character rules on this page describe Classic. The six science arcade games have their own controls and simplified physics.
THE TWO-GAP QUESTION
Before every input change, ask: “Where do I want to leave this opening?” The nearest gap decides whether you survive; the following gap decides whether that survival creates a usable position. A centered crossing is not automatically best. A slightly low exit may be ideal before a rising opening because it preserves room to hold.
READ BOTH POSITION AND DIRECTION
“The spider is low” is incomplete. Low and rising may require release; low and falling may require an early hold. Use the spider’s body angle and recent curve as direction cues. The visible web confirms held input, but its attachment point is not a target that steers the spider sideways.
MEASURED CASE: A SMALL DELAY DURING A FAST FALL
Here is one inspectable example of scenario 03. Both comparisons start from the same scripted mid-run snapshot: a 440 × 700 px portrait arena, spider center 340 px from the top, downward speed +300 px/s, and the next tree center 120 px ahead. The openings have centers at 360 and 280 px. Their 80 px change fits the generator’s portrait constraints; this hand-selected course is not a sampled random run.
Run A holds immediately. Run B waits six frames (0.100 seconds), then uses A’s remaining input schedule at the same absolute times. Every movement step uses Classic’s current acceleration, speed caps, collision box, scrolling speed, and update order. This is a deterministic scripted comparison at 60 steps per second, not a human recording or a browser performance test.
Drag the slider or use its arrow keys. Nothing plays automatically. The box is Classic’s collision rectangle; positive vertical speed means falling.
- Input for this step
- Hold
- Center from top
- 340.0 px
- Vertical speed
- +300.0 px/s
- Input for this step
- Release
- Center from top
- 340.0 px
- Vertical speed
- +300.0 px/s
| Run | Outcome | Smallest clearance while overlapping a tree |
|---|---|---|
| A · Immediate hold | 2 gaps cleared by 2.300 s | 59.25 px |
| B · 0.100 s delay | First-tree collision at 0.350 s; 0 gaps cleared | -1.47 px (overlap) |
The useful action is to cancel the fall while horizontal room remains. A’s initial hold lasts 23 frames (0.383 seconds). B has already used six of those frames falling before it starts correcting. By 0.350 seconds its collision box overlaps the lower tree. This example demonstrates sensitivity to entry momentum; it does not establish a universal reaction-time limit or an optimal control policy.
Exact inputs, assumptions, and reproduction files
A holds during frame intervals [0, 23), [40, 55), [73, 94), and [121, 138), and releases otherwise. B additionally releases frames [0, 6). Each interval ends before its second number. The manually chosen schedule illustrates a successful correction; it was selected after comparing candidate timings. It is not a measured distribution of typical player performance.
Height increases downward. Obstacles are 86 px wide; the spider collision half-width/height is 17/16 px. World scale is 700/680, scrolling speed is 185 × scale, opening height is 174 × scale, and obstacle spacing is 248 × scale. The simulation uses a fixed 1/60-second step; real Classic uses a variable animation-frame step capped at 0.034 seconds. This is a finite two-obstacle course without recycling or new trees. It includes Classic’s immediate crash-speed reset, but excludes rendering, resizing, input latency, and post-crash animation.
Download every frame (CSV), the setup, results and source hashes (JSON), the simulation, production rule constants, the production game source snapshot, and the reproduction script. Keep the last four files in their shown lib/, app/, and scripts/ directories. With Node 22.18 or newer, run node scripts/generate-strategy-evidence.mjs from their parent directory. Output is written under public/research/strategy/.
LOW ENTRY, LEVEL OR FALLING
BEGIN THE RECOVERY BEFORE THE TRUNK
Use a moderate hold while horizontal room remains. Watch for the curve to flatten; release before the spider reaches the gap center so remaining upward momentum carries it through instead of into the upper cap.
AVOID: A series of emergency micro-taps beside the trunk. Each may change speed, but the path has too little time to move.
HIGH ENTRY, STILL RISING
RELEASE EARLIER THAN LOOKS NATURAL
Stay released and let gravity cancel the rise. If the next gap is also high, use only a short stabilizing pulse after the upward speed has fallen. Preserve headroom above the spider.
AVOID: Holding merely because the gap center is above the spider; direction matters as much as position.
CENTERED, FALLING FAST
TRADE HEIGHT FOR A FLATTER EXIT
Start a hold immediately, not to climb at once but to reduce the downward speed. Accept a lower crossing if it gives the curve time to flatten before the following obstacle.
AVOID: Waiting for the spider to look low. At the +440 px/s cap, appearance lags behind the correction you already need.
SHARP HIGH-TO-LOW REVERSAL
THE FIRST GAP SETS UP THE SECOND
Cross the high opening below its center and already released. That converts remaining horizontal distance into the drop required by the lower opening. Add lift only after the fall is clearly established.
AVOID: Celebrating the first clearance with a long hold; it erases the planned descent.
EMERGENCY EDGE RECOVERY
CHOOSE THE SURVIVABLE EDGE
If perfect centering is gone, prioritize clearing the nearer cap by a small margin. Commit to one continuous correction, then prepare an immediate opposite input after the trunk.
AVOID: Alternating rapidly without reading velocity. Contradictory pulses consume distance while barely moving the path.
A CONTROLLED FIVE-RUN SESSION
- Use one run to identify direction aloud: “rising,” “level,” or “falling.”
- Use one run to cross every opening slightly below center.
- Use one run to look at the second gap before changing input.
- Use one run to recover with a single committed hold instead of rapid taps.
- Play normally and note which scenario still creates the collision.
A short focused session makes the input-response relationship easier to remember than endless undirected retries.
WHEN TO USE THE OTHER GUIDES
If the spider’s response still feels mysterious, the visual tutorial compares real paths and the physics guide provides the numbers. If the challenge seems to come from the generated sequence rather than your entry, inspect the constraints and limitations in the fairness report.