Bot Arena

The same 200 people and 5 sniper bots go through Sealround, a bonding curve and a first-come fixed-price sale, with the bots acting first.

Run a simulation

Sealround

Bot profit: 0USDC; 15.7% of supply bought
People paid: 1.00×; the bots' price

Bonding curve

Bot profit: 1,269USDC; 34.5% of supply bought
People paid: 2.61×; the bots' price

Fixed price, first come

Bot profit: 177USDC; 20.9% of supply bought
People paid: 1.00×; the bots' price

Simulated: averages over 60 seeded runs, with demand from 0.5× to 4× supply. Real launches will vary.

Results by launch type

The average of 60 runs, with the lowest and highest run under it.

Bot share of supplyLower is better
Sealround
15.7%
9.1 to 34.3%
Bonding curve
34.5%
28.7 to 43.1%
Fixed price, first come
20.9%
14.7 to 34.3%
Bot profitLower is better
Sealround
0 USDC
0 to 0
Bonding curve
1,269 USDC
700 to 2,068
Fixed price, first come
177 USDC
0 to 683
Human price premiumLower is better
Sealround
1.00×
1.00 to 1.00×
Bonding curve
2.61×
1.49 to 3.92×
Fixed price, first come
1.00×
1.00 to 1.00×
Humans filledHigher is better
Sealround
59.5%
37.0 to 84.0%
Bonding curve
50.1%
23.0 to 92.5%
Fixed price, first come
58.2%
28.0 to 85.0%
Holder concentrationLower is better
Sealround
0.50
0.41 to 0.64
Bonding curve
0.52
0.44 to 0.62
Fixed price, first come
0.50
0.39 to 0.64

Average of each metric over 60 runs, with the lowest and highest run below it.
Metric	Sealround	Bonding curve	Fixed price
Bot share of supplyLower is better	15.7%9.1 to 34.3%	34.5%28.7 to 43.1%	20.9%14.7 to 34.3%
Bot profitLower is better	0 USDC0 to 0	1,269 USDC700 to 2,068	177 USDC0 to 683
Human price premiumLower is better	1.00×1.00 to 1.00×	2.61×1.49 to 3.92×	1.00×1.00 to 1.00×
Humans filledHigher is better	59.5%37.0 to 84.0%	50.1%23.0 to 92.5%	58.2%28.0 to 85.0%
Holder concentrationLower is better	0.500.41 to 0.64	0.520.44 to 0.62	0.500.39 to 0.64

What each metric measures

Bot share of supply: Tokens bought by bots at launch ÷ all tokens bought at launch.
Bot profit: USDC bots receive selling to later human buyers, minus what they paid for the tokens they sold. Tokens they cannot sell count as zero.
Human price premium: Median price per token paid by humans who got tokens (launch plus pool) ÷ the average price bots paid at launch.
Humans filled: Share of the humans holding any tokens at the end, from the launch or the pool.
Holder concentration: Gini coefficient of final holdings across every account holding tokens. 0 = equal, 1 = one holder.

Each metric as demand rises

Tokens bought by bots at launch ÷ all tokens bought at launch.

SealroundBonding curveFixed priceOne run. Lines average 7 runs.

Demand: total human budget ÷ (supply × launch price)

All 60 runs as a table

Bot share of supply for every run
Run	Demand	Sealround	Bonding curve	Fixed price
1	0.50×	34.3%	43.1%	34.3%
2	0.56×	27.2%	36.4%	27.2%
3	0.62×	26.6%	39.3%	26.6%
4	0.68×	28.0%	39.8%	28.0%
5	0.74×	25.4%	38.4%	25.4%
6	0.80×	27.6%	39.8%	27.6%
7	0.86×	27.2%	41.8%	27.2%
8	0.92×	22.0%	36.9%	22.0%
9	0.97×	19.9%	36.1%	19.9%
10	1.03×	18.8%	35.5%	18.9%
11	1.09×	19.8%	36.5%	21.0%
12	1.15×	16.7%	33.0%	17.4%
13	1.21×	17.3%	34.9%	19.1%
14	1.27×	20.3%	38.1%	24.2%
15	1.33×	14.5%	31.8%	16.1%
16	1.39×	16.7%	35.0%	19.2%
17	1.45×	15.6%	33.6%	18.5%
18	1.51×	16.1%	35.0%	20.0%
19	1.57×	15.1%	33.8%	18.6%
20	1.63×	14.0%	33.2%	18.2%
21	1.69×	12.7%	30.6%	15.9%
22	1.75×	18.1%	37.6%	23.9%
23	1.81×	13.8%	32.8%	18.6%
24	1.86×	16.7%	35.1%	21.3%
25	1.92×	14.9%	35.1%	21.0%
26	1.98×	13.8%	33.3%	19.3%
27	2.04×	14.3%	34.5%	20.3%
28	2.10×	12.3%	32.5%	18.2%
29	2.16×	12.8%	33.1%	18.4%
30	2.22×	13.9%	33.8%	20.2%
31	2.28×	13.0%	33.3%	19.5%
32	2.34×	13.5%	33.7%	19.5%
33	2.40×	14.7%	35.0%	21.3%
34	2.46×	12.2%	32.0%	17.8%
35	2.52×	12.2%	32.6%	19.2%
36	2.58×	12.8%	33.5%	20.2%
37	2.64×	11.6%	31.6%	17.7%
38	2.69×	13.8%	35.6%	22.8%
39	2.75×	11.9%	32.8%	19.3%
40	2.81×	13.4%	34.9%	21.5%
41	2.87×	13.2%	33.4%	19.9%
42	2.93×	12.6%	31.8%	17.9%
43	2.99×	12.3%	33.9%	20.7%
44	3.05×	14.2%	34.3%	21.3%
45	3.11×	14.7%	34.3%	21.4%
46	3.17×	10.7%	31.1%	17.3%
47	3.23×	14.2%	35.7%	23.3%
48	3.29×	12.6%	34.4%	22.2%
49	3.35×	12.1%	33.9%	21.3%
50	3.41×	9.1%	28.7%	14.7%
51	3.47×	10.4%	31.8%	18.5%
52	3.53×	15.0%	35.8%	24.2%
53	3.58×	12.4%	33.6%	20.4%
54	3.64×	10.9%	32.0%	18.5%
55	3.70×	10.5%	32.4%	19.0%
56	3.76×	11.8%	33.5%	20.5%
57	3.82×	12.4%	34.8%	23.0%
58	3.88×	11.2%	32.8%	19.9%
59	3.94×	11.6%	33.2%	19.8%
60	4.00×	12.6%	34.7%	22.6%

Sealround: supply burned as demand changes

Share of the 1,000,000 tokens for sale that nobody bought. Sealround burns them when the creator is paid, so the supply that reaches the market follows demand. Simulate your own sale

Average of 7 runsOne run

Demand: total human budget ÷ (supply × launch price)

All 60 runs as a table

Sealround supply burned and clearing price for every run
Run	Demand	Burned	Clearing price
1	0.50×	38%	0.01 USDC
2	0.56×	42%	0.01 USDC
3	0.62×	26%	0.01 USDC
4	0.68×	24%	0.01 USDC
5	0.74×	22%	0.01 USDC
6	0.80×	21%	0.01 USDC
7	0.86×	3%	0.01 USDC
8	0.92×	11%	0.01 USDC
9	0.97×	2%	0.01 USDC
10	1.03×	0%	0.0101 USDC
11	1.09×	0%	0.0106 USDC
12	1.15×	0%	0.0105 USDC
13	1.21×	0%	0.011 USDC
14	1.27×	0%	0.012 USDC
15	1.33×	0%	0.0111 USDC
16	1.39×	0%	0.0115 USDC
17	1.45×	0%	0.0119 USDC
18	1.51×	0%	0.0124 USDC
19	1.57×	0%	0.0123 USDC
20	1.63×	0%	0.013 USDC
21	1.69×	0%	0.0125 USDC
22	1.75×	0%	0.0132 USDC
23	1.81×	0%	0.0134 USDC
24	1.86×	0%	0.0127 USDC
25	1.92×	0%	0.0141 USDC
26	1.98×	0%	0.014 USDC
27	2.04×	0%	0.0142 USDC
28	2.10×	0%	0.0148 USDC
29	2.16×	0%	0.0144 USDC
30	2.22×	0%	0.0145 USDC
31	2.28×	0%	0.0151 USDC
32	2.34×	0%	0.0145 USDC
33	2.40×	0%	0.0145 USDC
34	2.46×	0%	0.0145 USDC
35	2.52×	0%	0.0158 USDC
36	2.58×	0%	0.0159 USDC
37	2.64×	0%	0.0152 USDC
38	2.69×	0%	0.0165 USDC
39	2.75×	0%	0.0163 USDC
40	2.81×	0%	0.016 USDC
41	2.87×	0%	0.0151 USDC
42	2.93×	0%	0.0143 USDC
43	2.99×	0%	0.0169 USDC
44	3.05×	0%	0.015 USDC
45	3.11×	0%	0.0146 USDC
46	3.17×	0%	0.0162 USDC
47	3.23×	0%	0.0164 USDC
48	3.29×	0%	0.0176 USDC
49	3.35×	0%	0.0177 USDC
50	3.41×	0%	0.0161 USDC
51	3.47×	0%	0.0178 USDC
52	3.53×	0%	0.0162 USDC
53	3.58×	0%	0.0164 USDC
54	3.64×	0%	0.0169 USDC
55	3.70×	0%	0.0182 USDC
56	3.76×	0%	0.0173 USDC
57	3.82×	0%	0.0186 USDC
58	3.88×	0%	0.0179 USDC
59	3.94×	0%	0.017 USDC
60	4.00×	0%	0.018 USDC

Run one yourself

The same simulator runs in your browser. A seed and a demand level always give the same result.

Seed

Demand2.00× supply

Sealround, everyone: 0.0147USDC
Curve, bots: 0.00387USDC
Curve, median human: 0.00958USDC; 2.48× the bots' price
Fixed price sold out: at 8m 53s

Sealround clearing priceBonding curve priceFixed price, until it sells outBot buys at second 0

Time after launch, square-root scale. Price in USDC per token.

Bot share of supply
Sealround
11.6%
Bonding curve
31.4%
Fixed price, first come
17.0%
Bot profit
Sealround
0 USDC
Bonding curve
922 USDC
Fixed price, first come
177 USDC
Human price premium
Sealround
1.00×
Bonding curve
2.48×
Fixed price, first come
1.00×
Humans filled
Sealround
58.5%
Bonding curve
45.0%
Fixed price, first come
54.5%
Holder concentration
Sealround
0.50
Bonding curve
0.53
Fixed price, first come
0.53

Metrics for seed 20261114, demand 2.00×.
Metric	Sealround	Bonding curve	Fixed price
Bot share of supply	11.6%	31.4%	17.0%
Bot profit	0 USDC	922 USDC	177 USDC
Human price premium	1.00×	2.48×	1.00×
Humans filled	58.5%	45.0%	54.5%
Holder concentration	0.50	0.53	0.53

Recorded run: Lantern

A settled sale on this demo's simulated chain, sealed and cleared by the real code. Deposits and times were public during the window. Max prices stayed sealed.

Method and assumptions

pnpm arena

One command reruns all 60 runs and rewrites the data behind this page. The simulator is in packages/arena; its README.md covers the model.

Base seed: 20261114
Runs: 60, demand 0.5× to 4×
Crowd per run: 200 humans, 5 bots

Model assumptions 16

Written out from the simulator config on each run, so the list matches the code that made the charts.

Supply for sale: 1,000,000 tokens. Reference launch price: 0.01 USDC per token, so selling the whole supply at that price raises 10,000 USDC.
Demand level = total human budget ÷ (supply × launch price). 60 runs, evenly spaced from 0.5x to 4x. Each run draws a fresh crowd from its own seed, derived from base seed 20261114.
200 humans per run. Each has a private valuation, the most they would pay per token: lognormal, median 1.6x the launch price, sigma 0.55, clamped to 0.3x–12x. Valuations are never shown.
Human budgets are lognormal weights (sigma 0.9, so a few whales) scaled to the demand level, at least 1 USDC each. Humans arrive between second 1 and second 1,800, more of them early.
5 sniper bots per run, each with a budget of 2%–6% of the supply value (200–600 USDC). Bots act at second 0, before any human. On the bonding curve and in the fixed-price sale they buy with their whole budget. In Sealround they bid their whole budget with a max price of 100x the launch price.
Bonding curve: constant product x·y = k with virtual reserves in pump.fun proportions (virtual token reserve = 1.3529 × the curve supply). The virtual USDC reserve is set so buying the whole curve raises 10,000 USDC, the same as the fixed-price sale. The price runs from 0.002608 to 0.038335 USDC (0.26x to 3.83x the launch price).
On the curve, buyers act in arrival order. A human buys until their budget runs out or the price reaches their valuation. The curve is a two-way market from second 0, so after each human buy, bots sell back into it with the same rule as the post-launch pool (below). The launch ends when the curve sells out or the last human has arrived.
Fixed-price sale: 0.01 USDC per token, first come first served until supply runs out. A human buys only if their valuation is at least the price.
Sealround: every human whose valuation is at or above the floor (the launch price) bids max price = valuation and max spend = deposit = whole budget. Humans below the floor do not bid. Clearing uses clear() from @fairground/clearing, the same rule the enclave runs. Unsold tokens are burned.
Humans are honest and simple: they bid their true valuation in Sealround and stop at it on the curve. Nobody splits across wallets.
Post-launch market, identical for all three: a constant-product pool opens at the launch's final price (curve: its last price; fixed price: 0.01; Sealround: the clearing price) with 2,000 USDC of liquidity (20% of the supply value) paired with tokens at that price.
Humans with budget left and a valuation above the pool price buy from the pool in arrival order, each until their budget runs out or the price reaches their valuation.
Bots sell into that demand: after each human buy, bots sell back until the price returns to where it was before that buy. A bot never sells below what it paid per token. Tokens bots cannot sell stay in their wallets.
Bot profit counts only tokens actually sold, at the average price the bot paid. Unsold tokens count as zero, not at a market value. Bots follow this one simple selling rule everywhere; bots that time their sales better would likely make more on the curve and in the fixed-price sale, so those profits are conservative.
No fees, gas, latency or MEV beyond bots acting first. Money is integer throughout (USDC 6 decimals, tokens 18 decimals); floats appear only in the reported metrics.
Deterministic: the same seed and config give byte-identical output. No LLM calls, no network.

Watch it on a live demo sale. Judge mode runs a sniper and a copycat against the mock contract: the sniper pays the clearing price like every winner, and the copied bid is voided.

Open judge mode