Golden Bot vs. Bot Bot With Gilded Sniper: 10 Short Tests

Published 2026-11-01

My original Golden Bot and Bot Bot setup beat three alternative allocations in 10 short tests. More Golden Bot range raised its reported kill coverage, but coins per hour fell. Moving medals toward longer Golden Bot duration also failed to improve the one sample I tried.

That leaves me with a useful baseline and a better testing method. It does not establish the best duration or range for every tower, or even the best long-run allocation for mine.

If you have Gilded Sniper and limited medals, keep a working preset while you test. A higher coverage percentage or a plausible spreadsheet model is a reason to investigate a change, not enough evidence to replace a farming setup.

The setup behind these tests

These were Tier 19 starts on November 1, 2026. Each sample lasted about 10 minutes of game time, or 2 minutes 30 seconds to 2 minutes 33 seconds of real time, and ended at wave 20 or 21. They were deliberately short checks, not full farming runs.

The shared setup was:

• An Ancestral Gilded Sniper cannon.
• A displayed tower range of 43.5m, using the Range card with minimum workshop range. This was not a 30m tower.
• Full Golden Tower and Black Hole uptime, as configured on this account.
• Golden Bot at a 50s cooldown and an 8x bonus; Bot Bot at a 50s cooldown and a 2x bonus.
• No shared path through Synchronicity. Matching cooldowns did not make the bots travel together.

The duration values below are the final in-game values, including labs. Bot ranges are the displayed values in this setup. The samples were not randomized paired trials, and activation phase and individual kill locations were not measured.

Four allocations and all 10 results

CPH means coins per real-time hour. The game's q means quadrillion coins; T means trillion. The mean below is the arithmetic mean of the CPH values shown in the reports, not a forecast for a full farm.

The individual reported CPH values were:

• A: 4.31q, 3.78q, and 4.18q.
• B: 4.04q, 3.62q, and 3.55q.
• C: 3.97q, 3.72q, and 3.30q.
• D: 3.34q, from one run only.

Relative to A's sample mean, B was down 8.6%, C was down 10.4%, and D was down 18.3%. Those are observed differences between these samples. They are not established losses caused by the allocations. In particular, B and C differ by only about 2%, while C's own samples span 3.30q to 3.97q.

The alternatives also left slightly different amounts of medals unspent. These were practical allocation trials, not an equal-cost measurement of one upgrade's marginal return. You can inspect the selected report data and calculation notes.

Higher Golden Bot coverage did not mean higher income

The battle report's Golden Bot entry under Killed With Effect Active rose as I expanded its range. These percentages use the summed counts for each setup, rather than an average of rounded report percentages.

The last column is a diagnostic I calculated by dividing the report's Other Coin Bonuses field by total enemies. It is not the game's similarly named Coins / Kill field, a base enemy coin value, or an independently attributable bot contribution. B means billion coins. Coin-source fields overlap, so adding the Golden Bot, Black Hole, and Golden Tower entries would double-count income.

Setup C is the clearest warning about optimizing coverage alone. It recorded more GB-active kills than A across three runs, with nearly the same total enemy count, yet its CPH was lower. Setup D had nearly the same GB-active percentage as A and still produced a lower result.

The report does not show whether those were the same-value enemies, the same multiplier combinations, or the same Golden Combo timing. It also does not directly reveal coin-weighted overlap between Golden Bot and Bot Bot. A rising GB percentage cannot answer those questions.

Gilded Sniper changes the trade, without removing range

The official Gilded Sniper explanation gives Ancestral a 40% proc chance and identifies the eligible coin effects. It explicitly excludes Bot Bot from the effects GS applies. The original Bot Bot announcement describes its effect as operating within its radius; Synchronicity is a separate pathing upgrade.

For a covered death, active Golden Bot supplies its normal bonus without relying on GS. For an otherwise eligible death outside its range, GS offers a chance to recover that bonus. Under a simplified GB-only model, effective coverage during the active window is:

effective coverage = c + p * (1 - c)

c = physical coverage of eligible deaths while GB is active
p = Gilded Sniper proc chance

At p = 0.40, that becomes 0.40 + 0.60 * c. Increasing physical coverage still helps. This model assumes comparable coin value across those deaths and excludes Bot Bot, other multiplier correlations, and Golden Combo feedback. A bot's circle area is not a measurement of c.

October Discord discussions describe GS-awarded GB bonuses being amplified when the death is physically inside Bot Bot. That is a player-reported interaction, separate from claiming that GS applies BB everywhere. The October 20 discussion and October 28 diagram require server access. These tests do not isolate or verify that interaction, so I keep both interpretations available in the calculator.

Setup C reduced BB range from 85m to 75m to fund more GB range. Its lower income is consistent with losing valuable BB overlap, but it does not prove that explanation. The bots had independent paths, and we did not measure their joint coverage. Setup D changed all four duration and range values at once, so it cannot isolate duration either.

Players comparing GS respecs on Reddit also reported mixed outcomes, including losses after sacrificing range and a case where a respec disturbed Summon timing. Those accounts help identify variables to control; they do not provide a universal allocation to copy.

What the 52% duration threshold actually means

For my original setup, the next GB duration purchase cost 980 medals and added 0.5s to 41s. BB's next duration purchase cost 580 medals and added 0.5s to 36s.

Suppose repeated, comparable tests give three income measurements with GS equipped throughout:

• C0: both bots disabled.
• CG: Golden Bot enabled and Bot Bot disabled.
• CGB: both bots enabled.

Define A = CG - C0 as GB's incremental income in that comparison, and B = CGB - CG as BB's additional income with GB already enabled. If those increments scale proportionally with added duration, BB's next duration purchase wins per medal when:

B / A > (36 * 580) / (41 * 980)
B / A > 0.5197

That is the source of the roughly 52% crossover. It is specific to those durations and next purchase prices. It is not BB's share of total coins or a coverage percentage in the report.

All 10 runs above had both bots enabled. They do not supply the three on/off measurements, so they cannot tell us whether BB clears that threshold. Even with those measurements, the proportional-duration assumption can fail if the added half-second misses the valuable kill window, changes overlap, or interacts with hoarding and payout timing. The calculator's estimate should be paused when those assumptions do not fit the run.

For choosing between complete presets, direct repeated A/B comparisons are usually more useful than collecting bot-off runs just to estimate this threshold. Bot-off tests also give up farming income while they run.

How I would test the next change

My next useful control is a return to A under the same conditions. A is the best observed baseline here; its 41s GB duration has not been proven optimal.

1. Repeat the baseline before adding another allocation. If A also falls well below its earlier results, investigate normal variation, run boundaries, and unintended setup changes first.
2. Compare one factor at a time where the medal budget permits. If a fixed-budget trade requires changing two stats, record both and evaluate the complete trade. Keep cards, modules, tower range, labs, speed, and activation procedure consistent.
3. Alternate configurations. Use an order such as A, B, B, A, then repeat. Record every result rather than keeping only the best run.
4. Use matching farming windows. For a stable section, measure the change in coins divided by elapsed real time. Start and end at comparable points relative to Golden Combo payouts; avoid comparing a settled payout with a still-pending one.
5. Confirm promising changes in normal farms. A wave-21 start does not represent the enemy mix, perks, survival, and income pattern later in a run. Track cells and shards too if they matter to your farming goal.

A larger sample helps, but there is no fixed number of short runs that makes every result conclusive. When the observed difference is small compared with run-to-run variation, keep the conclusion open.

Use the calculator with the right inputs

The Bot ROI Calculator now separates the next upgrade from a full medal-budget allocation. Enter the final displayed bot values and the actual next medal prices. Its current-setup mode accepts repeated CPH observations, keeps GS rarity explicit, and distinguishes matching cooldowns from shared paths.

If you use its coverage model, supply assumptions you can defend. Unknown joint coverage produces a range of estimates. Range upgrades stay unranked until you provide an assumed or measured coverage gain, and uncertain GS/BB behavior can be compared under both interpretations. The older budget planner's area score remains a geometry aid rather than a measured CPH forecast.

Use the Battle Report Analyzer to retain reports and setup notes, the Bot Simulator for upgrade planning, and the earlier Gilded Sniper article for the module overview. My current decision is to keep the original farming baseline and collect a cleaner comparison before reallocating more medals.