Twitxy Saved

Brwn 🔮

Verified account This account is verified. Learn more

@brwnpb

enabling algorithmic + automated trading via @useOccam
t.me/OccamPlatform Joined September 2023
810 Following 2.5K Followers 1.8K Posts

Latest posts 5

Pinned post

Brwn 🔮 Verified account @brwnpb
Time to announce what's next.

We're bringing algorithmic & automated trading to the trenches through @useOccam offering:

> Algorithm Creation
> Algorithm Marketplace
> Automated Trading

Enabling onchain market exposure to those with jobs or sleep (99.9% of traders btw)
12 replies 16 reposts 71 likes 19K views View on X
Brwn 🔮 Verified account @brwnpb
Patch Notes - Last 72h:

🫂Fixed - Reported Issues:

- 7D simulations that include bonding-curve tokens now finish instead of failing
- Simulations that start or end part-way through a day now run properly
- Long simulations over older dates (100+ days / 36,000+ tokens) are split up behind the scenes and finish, instead of being refused
- If the system is busy, your simulation now waits its turn instead of failing. If a run ever does fail, the message tells you it's safe to submit again
- The rule finder now works on date ranges that end part-way through a day
- Token names now show on very large bonding-curve simulations
- The bonding-curve badge in the algorithm builder only appears when "include curve tokens" is switched on

⚡️Speed:

- 100+ day backtests over older dates run 4x faster
- Re-running or tweaking a long backtest run 7x faster
- 2-week backtests over older dates run 3x faster
- Large bonding-curve backtests now run 11% faster, and no longer stuck retrying. Work in progress
- Standard backtests no longer wait behind bonding-curve runs
- Typical times today: 1 week in 53 s, 1 month in about 2 min, 2 months in about 4 min

⚙️Accuracy:

- Peak prices on older dates are more accurate. Brief price spikes no real trade could have caught are no longer counted, which corrected peaks that were overstated by up to 5x (for example 9.9x is now correctly 2.1x)
- Fraud metrics now cover their full history, not just the last 7 days
- Initial Bundle SOL is now recorded from 8 October onwards
- The metrics panel shows the earliest date each metric can be simulated from

🔒Reliability:

- 93% fewer database errors
- Token labels now update steadily through the day

🛣️Still being worked on:

- 26 October, 20:00–24:00 UTC is missing most token data, so results for that evening read low. We're working on filling it
- Occasional "please try again" errors at busy times. A database improvement is rolling out tonight to remove the last of these
- Brand new long backtests over older dates can take 5–7 minutes the first time, repeats and tweaks are much faster
- Re-running a bonding-curve backtest will reuse earlier work, improving speed by about 4-5x
1 replies 9 reposts 29 likes 774 views View on X
Brwn 🔮 Verified account @brwnpb
Gunna solely focus on non-technical stuff here ahead of the large feature update tomorrow

User Experience, Profitability & Misc:

User Experience🫂

We've encouraged the API to feed things to you concisely, without the technical fluff & give headlines:

> You see the answer straight away, e.g did it make money, and how often does it trade.

> You get a short reply instead of a wall of numbers.

> Long tests tell you how long they'll take

> All your past tests are in one list, so nothing gets lost

> We feed example prompts to take you down the right path

> Very large tests no longer break. You're told how to shrink them instead or we do it ourselves for you.

Profitability ⚖️

> We visualise if your algorithm actually beats picking tokens at random.

> We visualise and make clear which multiples suit you best for trading too

> Full clarity on overfitting and encouraging analysis away from it

> Optimal sell timing

> You can put up to four algorithms side by side and back the one that earns most

> Older results are now correct, so you can trust long-term tests.

> Improvements on searching for a winning algorithm for you.

Misc Improvements🎲

> Opens every answer with the bottom line (profit, trades per week, verdict)

> Users are updated throughout analysis and time left on long tests.

> A list of all past tests, and a paper-mode overview.

> Old history loads faster.

> Week-by-week results.

> A built-in guide to hunting for a winning algorithm.

> Updated documentation & Daily Changelog on API updates

> You no longer have to reconnect to the platform to receive latest API updates, it installs automatically.
Brwn 🔮 Verified account @brwnpb Nov 5
honestly, I don't care if markets go up or down

all I care about is volume, there's opportunities in that

there's about +$500m daily vol

until we've captured a large portion of that w/ @useOccam & solana:2sQ7wuUtRWNir3CEu9HWfLDSut4AszDrcZXLobzJpump then there's nothing to complain about

next update out tonight, brb
View quoted post on X
2 replies 8 reposts 29 likes 759 views View on X
Brwn 🔮 Verified account @brwnpb
honestly, I don't care if markets go up or down

all I care about is volume, there's opportunities in that

there's about +$500m daily vol

until we've captured a large portion of that w/ @useOccam & solana:2sQ7wuUtRWNir3CEu9HWfLDSut4AszDrcZXLobzJpump then there's nothing to complain about

next update out tonight, brb
2 replies 8 reposts 32 likes 1.3K views View on X