Using Materral

Paper trading. Test a gap against the live market.

A paper trade replays a route's legs against the live quotes after each leg's latency, then compares the realised result with what Materral expected. No order is sent anywhere.

The trade ticket#

Size#
Start from a gap and confirm the size. The ticket prices both legs at that size from the current books and pools.
Review#
Check each leg's expected VWAP, the limit price (expected VWAP ± max slippage), the cost lines and the expected net. Max slippage defaults to 10 bps.
Place#
Place the paper trade. Each leg is submitted at once and fills after its leg latency, 2 s by default (an assumption). The trade then settles into one of the outcomes below. A route with a blocking flag, such as a stale quote or a down feed, cannot be placed.

The fill model#

The model is deliberately conservative.

Each leg is an immediate-or-cancel order with a limit at the expected VWAP ± max slippage.
It fills after the leg latency, against the live quote current at that moment, not the quote you saw.
Book legs fill level by level up to the limit; the rest is cancelled, so a partial fill is possible.
If the depth snapshot is older than the best bid/offer, that side's levels are shifted by the drift first.
Pool legs fill fully or revert, like a swap with a minimum output. That includes size beyond the current tick range.
The legs are separate, non-atomic orders. One can fill while the other fails, leaving you exposed.

What can fail, and why#

Reason	What happened
`no_fresh_quote`	The quote was stale or missing at fill time. A leg never fills against old prices.
`feed_down`	The venue's feed was down at fill time.
`slippage`	The price moved past the limit.
`exceeds_tick`	A pool leg needed more size than the loaded tick range could price, so it reverted.
`no_liquidity`	Nothing could be filled.
`interrupted`	The leg was still unfilled 15 s after it was due, usually because the page was closed or paused.

Outcomes#

Complete: both legs filled.
Legged out: the legs did not fill the same amount (one failed, or one filled only partly), leaving residual exposure. It can be unwound at market.
Failed: nothing filled.
Closed: a legged-out trade that was unwound back to flat.

Reading expected vs realised#

Expected net is what the route showed when you placed it, including the latency-risk charge.
Expected cash is the same net before the latency-risk charge. That charge is a risk allowance, not money spent, so this is what realised is compared with.
Realised net is the USDC result of every filled leg, its fees and gas. Leftover inventory is valued at mid and shown separately as residual; it is marked, not realised.
Error is realised minus expected cash. Slippage per leg is realised VWAP versus expected, against you (positive is worse).
Pool fees are inside the realised VWAP, not in the fee line.

What the model cannot see#

your queue position;
hidden liquidity;
other traders reacting to your order;
gas price spikes between decision and fill;
real transfer times between venues (rebalancing is not modelled).

Paper trades are kept in your browser's local storage. Every one is labelled PAPER.

materral docs. Read-only, no keys, no real orders. Not investment advice.