PAL

Autonomous trading infrastructure for prediction markets. Live on @Polymarket.

[PAL @polyagentleague]

We've officially broken into the top 20 @PolymarketBuild by volume! In just 4 days, we've almost done $150k in volume with our agents.

Still a long way to go but all about optimising and improving the agents. We're coming for that top 10 spot.

For those of you new to PAL, make sure to join the community telegram - https://t.co/kBMMuvex4S

[PAL @polyagentleague]

New Agent Strategy Deployment: Contrarian - Beta Testing Phase underway.

Overnight, one of @polyagentleague Contrarian Agents just captured a 100x outcome on @Polymarket. Although exciting, for us, what makes that interesting is not the return itself, but that the behaviour behind it was not accidental.

Contrarian is designed to detect short-lived dislocations in fast-moving markets, then take small positions where downside is tightly bounded but upside can be highly asymmetric.

This Agent had been alive for a few hours before landing this trade, and it had already executed roughly 36 small $1–$2 entries around the same general setup. We noticed multiple positions it was taking came very close to resolving correctly (within 3-4$). So this was not a fluke, but rather an agent repeatedly expressing the same idea with discipline until the distribution paid - and it did.

If these behaviours continue to hold in live conditions, the implication is much bigger than one profitable outcome. It points toward agents that can 24/7 systematically detect and act on micro-opportunities that are structurally difficult to exploit manually.

That is what we are building at PAL.

Still in testing. Still early. But this is the kind of behaviour worth paying attention to. @PolymarketDevs

[PAL @polyagentleague]

@ragebeloved Can test and deploy the scalper/speed demon agent here: https://t.co/q9TMGFUdGF

If you have specific questions or need trouble shooting, the most efficient way is to join the community tg: https://t.co/3nHyxv3zc4

[PAL @polyagentleague]

This may look unassuming at first, but if you look closer, you can see exactly what PAL is doing.

This agent was funded with just $25 and plugged into our EdgeLab infra. It’s now officially positive all-in costs!

In a matter of days, it has traded over $7.5k in volume and is sitting in the green.

That’s what is exciting to us.

Given how PAL is built, and the sheer amount of trading this agent is already pushing through, this is not some fluke. It is doing exactly what we wanted it to do. We are not just turning on automation and praying it works. This is deliberate, structured, and already behaving the way we expected. That is the exciting part, because we built PAL to scale.....

And this is just one of our strategies.

We’ve opened PAL up to some beta testers now, with vaults next on the menu 👀 @PolymarketDevs @Polymarket

[PAL @polyagentleague]

In the last 24 hours the @polyagentleague community have started deploying multiple new scalper/speed demon agents. 'Polymaxxing' is one example, and already in it's first day it's at +$8.56 net after all costs on $3,071 volume on hundreds of trades (Max trade size was 30$). Volume and profit whilst you sleep. This is no accident. Full stack autonomous @polymarket trading. No manual intervention and everything is verifiable on-chain. In fact, each agent has a command centre and a "neural link" that shows the agents actions in real time.

We at PAL are extremely excited about what we are building and still only getting started. We built to scale. Come and see it for yourself....

[Juzz @Urqkel]

Started my poly agent to test for bugs

Set up was smooth, funded the agent with SOL and bridged perfectly, within an hour it made it first trade

Current trades below and this is real not like the other bs you see on CT

Might be time you had a closer look at $PAL

[PAL @polyagentleague]

Nice to see people that get it! All in per $ costs matter and of course, does it scale? Watch us get there!

[PAL @polyagentleague]

https://t.co/VdSyqGUISL

[PAL @polyagentleague]

@sol_goldrush @Polymarket @openclaw Just getting started

[PAL @polyagentleague]

EdgeLAB testing update:

Our @Polymarket agent was already around ~50 bps all in execution cost per $ traded. This week we’re now consistently at ~20 bps. 125 trades and $5k volume on limited maker orders only. That's a 60% cut with EdgeLAB & @openclaw working together.

At majority maker flow, this should tip net positive...👀

If you understand microstructure, you understand how big this is.

[PAL @polyagentleague]

Stop shipping random @Polymarket agents and pretending your wrapped LLM is gonna be the edge. If you can’t beat fees, spread, slippage, and compute at real volume, you’re lying to yourself. Most of the timeline is simply engagement farming.

PAL already trades Polymarket 24/7 with deterministic execution and layered risk controls. Our cost of volume is already ~0.5%.

Our next step is Edge Lab with @openclaw integrated: automated experiments, audit grade cost breakdowns, controlled paper and shadow cohorts, and promotion gates that only ship net positive upgrades as auto PRs.

We started around almost neutral cost of volume and now are pushing to be net positive before anyone else. Watch.

[PAL @polyagentleague]

https://t.co/OJnksfJbCQ

[PAL @polyagentleague]

I know comms could be better but last 1-2 weeks have been tough.

I've been working in last few days and should be fully back from tomorrow. Plan is to test and fully release the new updated agents within a week.

Will most some more technical details tomorrow.

[PAL @polyagentleague]

My apologies for lack of updates in last few days. Had a family emergency that I had to tend to so was out of the country. I'm back now and will get some updates out very soon.

Looking into HIP-4 to see what we can do there and then focus is back on the ML agent I was developing.

[PAL @polyagentleague]

My apologies for lack of updates in last few days. Had a family emergency that I had to tend to so was out of the country. I'm back now and will get some updates out very soon.

Looking into HIP-4 to see what we can do there and then focus is back on the ML agent I was developing.

[PAL @polyagentleague]

We're building an ML agent to trade Polymarket's 15-minute crypto markets.

Hundreds of these pop up daily, they resolve against live Chainlink oracle price data, and most traders ignore them entirely.

We already have a scalper agent running on Polymarket with a -0.5% return on volume, which in prediction markets is actually solid. But scaling a rule-based scalper only gets you so far. To find deeper edges, we need machine learning.

So we're doing it properly.

Step 1: Data collection. Our agent runs 24/7 in shadow mode - scanning every market, capturing orderbook depth, pulling live prices, and logging it all alongside the actual outcome. Building the labeled dataset the model needs to learn from.

Step 2: Paper trading. Simulate real trades against live markets. Track P&L, win rate, and calibration across thousands of predictions. No real money. This is where we're at right now.

Step 3: Train the model. Use the labeled data to learn which conditions lead to mispricing - when the Polymarket odds don't match what the exchange price says.

Step 4: Real money as long as the paper numbers hold, with strict risk limits.

When we figure this out, and we will, it will basically be unlimited volume at no cost.

[PAL @polyagentleague]

Our agents' performance is trending the right way and win rate is slowly improving. I've been fine tuning the settings and actively testing to try and get cost of volume even lower. So far so good.

Also, designing the machine learning aspect of it and feeding it thousands of past trades so it can learn from a bigger sample.

[PAL @polyagentleague]

A little introduction to PAL

[PAL @polyagentleague]

Have spent the last 2 days improving the scalper and trend grinder agents. Focusing on changing the market orders to limit orders for the scalper to get our cost down for volume. It's going pretty well and I think we should get some positive results soon.

On another note, I'll be making some videos from tomorrow to showcase the platform a bit more, how you can deploy an agent, fund it and so on.

[PAL @polyagentleague]

The new website is now live. I've completely revamped it to be more gamified and feel like a real game rather than a boring 'agent deployment' site. Some pics attached.

Check it out, test and deploy an agent. Let me know any feedback.

Time to now work on the trend grinder agent which has improved a lot but needs a bit more tuning.