Does hiring an agent on your marketplace beat doing the job yourself, and can you prove it?
6 tasks, each done both ways. Only policy= differs between the columns — same driver, same tape, same cost model, held fixed by construction.
Across equities, security, trading, on 4 venues.
COUNTERFACTUAL — neither position was heldchain tape
Across all 6 tasks
no verdict (5 observations, need 30)
6 tapes are not evidence about a strategy. Each task’s own quote is, and it is drawn from many windows.
tasks
6
quotable
5
withheld
1
agent ahead
2
DIY ahead
1
too close to call
2
bands separated
3
The 6 tasks
The baseline and the tape both differ per task. 4 on 31.0 days, 1 on 17.0 days, 1 on 4.1 days. Each column is compared against the other on that task’s own tape, so the window bounds how much of a year the annualised figure describes — not which of the two is ahead.
mint once at the same width, never touch it (passive_policy)8.94% – 16.35%
3 of 60 observations fall within 0.25% of the median. Every observation is above zero, which is off this axis.The bands are separated — the gap survives the spread, not just the medians. +17.21pp at the median.
+17.21ppagent beats DIY by 17.21pp, bands do not overlap
fees earned: 0.0578309 vs 0.0156938 — 3.7x better
material · bands separated · 31.0 days of tape
Won on
60 of 60 windows against doing it yourself
Decisions
1you make · 20made for you
Fee to hire
0.000755 BNBgas either way: 0.0008 DIY · 0.0155 agent
mint once at the same width, never touch it (passive_policy)8.94% – 16.35%
2 of 60 observations fall within 0.32% of the median. Every observation is above zero, which is off this axis.The bands are separated — the gap survives the spread, not just the medians. +12.71pp at the median.
+12.71ppagent beats DIY by 12.71pp, bands do not overlap
time in range: 0.999975 vs 0.366401 — 2.7x better
material · bands separated · 31.0 days of tape
Won on
56 of 60 windows against doing it yourself
Decisions
1you make · 13made for you
Fee to hire
0.000755 BNBgas either way: 0.0008 DIY · 0.0101 agent
Sentinel — withdraw on §3.4 toxicity, re-enter after m_clear-55.19% – -52.84%
the same wide band, held through everything (withdrawal disabled)9.39% – 14.94%
6 of 60 observations fall within 0.13% of the median.The bands are separated — the gap survives the spread, not just the medians. −68.55pp at the median.
−68.55ppagent loses to DIY by 68.55pp, bands do not overlap
realized convexity cost (upper bound on LVR): 0.00479899 vs 0.00854626 — 1.8x better
material · bands separated · 31.0 days of tape
Won on
0 of 60 windows against doing it yourself
Decisions
2you make · 1,191made for you
Fee to hire
0.000755 BNBgas either way: 0.0015 DIY · 0.0537 agent
Protect — avoid being picked off by one-way flow: agent against baseline
Metric
Sentinel — withdraw on §3.4 toxicity, re-enter after m_clear
the same wide band, held through everything (withdrawal disabled)
median return
-54.45%
14.10%
P25 – P75
-55.19% – -52.84%
9.39% – 14.94%
in range
87.0%
100.0%
fees (WBNB)
0.017
0.019
adverse selection (upper bound) (WBNB)
0.0048
0.0085
costs (WBNB)
0.054
0.0015
moves
1,191
2
net return on capital, P25–P75 (the cost of the withdrawals is charged in full) · on 1.00 WBNB of capital · annualised
security · two venues: PancakeSwap v3 WBNB/USDT 0.05% · 0x36696169C63e42cd08ce11f5deeBbCeBae652050 vs PancakeSwap v3 WBNB/USDT 0.25% · 0x1401ff943D08a7E098328C1d3a9d388923B115D2 — depth and the flow screen chose the same one
Choose — which pool to provide liquidity to
—±0.00pp
Without an agent
pick the deepest pool — PancakeSwap v3 WBNB/USDT 0.05% · 0x36696169C63e42cd08ce11f5deeBbCeBae652050, 84.0x the median liquidity (the obvious heuristic: more TVL is safer)
With an agent
pick the pool whose flow is not one-way — PancakeSwap v3 WBNB/USDT 0.05% · 0x36696169C63e42cd08ce11f5deeBbCeBae652050, where §3.4's imbalance arm fires on 41.9% of samples against 42.9% on the other — which is the pool depth chose too
-3.3%22.0%
pick the pool whose flow is not one-way — PancakeSwap v3 WBNB/USDT 0.05% · 0x36696169C63e42cd08ce11f5deeBbCeBae652050, where §3.4's imbalance arm fires on 41.9% of samples against 42.9% on the other — which is the pool depth chose too5.84% – 16.52%
6 of 60 observations fall within 0.22% of the median.
One band, because one thing was measured. Both rules chose the same venue, so the two columns are the same replay — same tape, same policy, same capital — and the difference between them is exactly zero by construction rather than by measurement.
±0.00ppindistinguishable: +0.00pp is below the 0.10pp materiality floor
immaterial · bands overlapping · 31.0 days of tape
Won on
one replay printed in both columns, so every window ties by construction
Decisions
15you make · 15made for you
Fee to hire
0.000755 BNBgas either way: 0.0008 DIY · 0.0008 agent
Choose — which pool to provide liquidity to: agent against baseline
Metric
pick the pool whose flow is not one-way — PancakeSwap v3 WBNB/USDT 0.05% · 0x36696169C63e42cd08ce11f5deeBbCeBae652050, where §3.4's imbalance arm fires on 41.9% of samples against 42.9% on the other — which is the pool depth chose too
pick the deepest pool — PancakeSwap v3 WBNB/USDT 0.05% · 0x36696169C63e42cd08ce11f5deeBbCeBae652050, 84.0x the median liquidity (the obvious heuristic: more TVL is safer)
median return
7.22%
7.22%
P25 – P75
5.84% – 16.52%
5.84% – 16.52%
in range
98.4%
98.4%
fees (WBNB)
0.0017
0.0017
adverse selection (upper bound) (WBNB)
0.00073
0.00073
costs (WBNB)
0.00081
0.00081
moves
15
15
net return on capital of the same agent on the chosen venue, P25–P75. Capital for this task is 0.03181, not the report's 1. A1's ceiling is eps x pool liquidity and therefore a property of the pool, and the shallower venue binds it — PancakeSwap v3 WBNB/USDT 0.05% · 0x36696169C63e42cd08ce11f5deeBbCeBae652050: 7.031 · PancakeSwap v3 WBNB/USDT 0.25% · 0x1401ff943D08a7E098328C1d3a9d388923B115D2: 0.06362 (at 0.5x margin). Both columns run at the same figure, because a comparison between two venues at two capitals is partly a comparison between two position sizes. · on 0.032 WBNB of capital · annualised
Quote withheld. 85 swaps spanning 4.1 days on PancakeSwap v3 TSLAx/USDT 0.25%. Assumption A5 requires 20 replay windows of at least 24h; this tape yields at most 4. It is the only tokenized-equity pool on BNB Chain with any liquidity — NVDAx and AAPLx have no v3 pool at any fee tier, and the 1.00% TSLAx pool has none. So the venue is real, our engine prices it, and there is not enough flow through it to quote. Withheld rather than estimated.
—no verdict — 85 swaps spanning 4.1 days on PancakeSwap v3 TSLAx/USDT 0.25%. Assumption A5 requires 20 replay windows of at least 24h; this tape yields at most 4. It is the only tokenized-equity pool on BNB Chain with any liquidity — NVDAx and AAPLx have no v3 pool at any fee tier, and the 1.00% TSLAx pool has none. So the venue is real, our engine prices it, and there is not enough flow through it to quote. Withheld rather than estimated.
Router - move only when the rate edge clears the round-trip cost1.71% – 1.92%
supply to the highest-rate venue once and never move (park_policy)1.90% – 1.92%
2 of 20 observations fall within 0.00% of the median. Every observation is above zero, which is off this axis.The bands overlap — at this sample size the two are not distinguishable, whatever the gap between their medians.
−0.01ppindistinguishable: -0.01pp is below the 0.10pp materiality floor
moves made (best rate seen 2.98%, hurdle 1.06%, break-even at 2 days): 2 vs 1 — worse
immaterial · bands overlapping · 17.0 days of tape
Won on
0 of 20 windows against doing it yourself · 13 tied
Decisions
1you make · 2made for you
Fee to hire
0.000755 BNBgas either way: 0.0152 DIY · 1.0304 agent
Route — which lending venue to supply to: agent against baseline
Metric
Router - move only when the rate edge clears the round-trip cost
supply to the highest-rate venue once and never move (park_policy)
median return
1.90%
1.91%
P25 – P75
1.71% – 1.92%
1.90% – 1.92%
in range
—
—
fees (WBNB)
—
—
adverse selection (upper bound) (WBNB)
—
—
costs (WBNB)
1.03
0.015
moves
2
1
Rows reading — are quantities this venue does not have — not measurements of zero. A lending market has no range to be in and pays no swap fee to a provider. This task asks only about lending venues, which is what its name says; Router’s own card asks the wider question and puts PancakeSwap ranges beside them.
net return on supplied capital (realized yield - switch costs), P25-P75, over a 168h stated holding period. Capital for this task is $10,000 of stablecoin, not the report's figure — that one is denominated in WBNB and means nothing to a lending market. · on 10,000.00 WBNB of capital · net return on supplied capital (realized yield − switch costs)
The same report, on too little history
The same code path on a tape too short to reach the policy horizon. It produces no numbers at all, which is the correct answer.
Deliberately short tapesynthetic tape4 of 4 withheld
Earn — fees on a liquidity position
no verdict — baseline: 0 usable replays, assumption A5 requires 20; 60 window(s) were shorter than the 24h policy horizon; agent: 0 usable replays, assumption A5 requires 20; 60 window(s) were shorter than the 24h policy horizon
Market-make — quote both sides of a range
no verdict — baseline: 0 usable replays, assumption A5 requires 20; 60 window(s) were shorter than the 24h policy horizon; agent: 0 usable replays, assumption A5 requires 20; 60 window(s) were shorter than the 24h policy horizon
Protect — avoid being picked off by one-way flow
no verdict — baseline: 0 usable replays, assumption A5 requires 20; 60 window(s) were shorter than the 24h policy horizon; agent: 0 usable replays, assumption A5 requires 20; 60 window(s) were shorter than the 24h policy horizon
Choose — which pool to provide liquidity to
no verdict — baseline: 0 usable replays, assumption A5 requires 20; 60 window(s) were shorter than the 24h policy horizon; agent: 0 usable replays, assumption A5 requires 20; 60 window(s) were shorter than the 24h policy horizon