Per-order minimums eat small size alive
i average maybe 12 round trips a day at like 800-1200 notional each, which is small on purpose while i test the entries. that costing study on commissions…
1
20d
TopicCategoryRepliesActivity
Why do my stops fill worse on the fast drops?
Been logging every exit for about six weeks now, mostly small caps off the premarket scan. Pattern I keep seeing: normal drift down, my stop fills basicall…
3
23d
The quote is a quantity, not a price
Been sizing up slowly on my 5m breakouts and kept noticing that my fills on the bigger clips sit worse than the small ones, same setup, same trigger, same…
4
25d
one day of sample data and my brain already wants to overfit it
So I grabbed that one-day Coinbase sample file from the downloads section, the one-minute one with the fifty-odd pairs, dropped it in the crypto folder, re…
1
29d
the tick loading writeup — worth it for a small symbol list?
morning all, coffee number two in hand ☕
3
Aug 2026
stale kraken tickers - how are you guarding against this?
read that research post on this site about kraken ticker updates going quiet on thin pairs and honestly it bugged me more than i expected.
3
Aug 2026
Building the normalized volume features from raw data — what I got wrong
Tried to build the volume ratio and turnover measures people described here. Both were harder than expected and I got them wrong in instructive ways.
3
Aug 2026
Survivorship, and the names that are not in your data
Everyone knows the term. Fewer people have actually looked at what's missing from their own dataset.
2
Aug 2026
What I'd want to see before believing any research posted here
We've had a productive stretch of posts and I'd like to propose a small convention, offered as a suggestion rather than a rule.
3
Aug 2026
Keep a data dictionary for your own factors
Small habit, disproportionate payoff. For every factor I compute, I keep three lines: the exact formula, the denominator or baseline it's normalized agains…
2
Aug 2026
Public flow data tells you what happened, never who
Some markets publish end-of-day summaries showing which brokers accounted for the largest share of a name's volume. Wherever that data exists, an entire co…
3
Aug 2026
1-minute candles can distort scalping backtests — a measurement
We compared the same scalping strategy on 1-second recordings vs 1-minute bars over the same days. The 1-minute version both missed entries (intra-minute t…
3
Aug 2026
Where your minute bars actually start
Small thing with consequences. When your data provider builds a one-minute bar, does it cover 09:30:00.000 through 09:30:59.999, or does it label that bar…
2
Aug 2026
Corporate actions are data quality events, not accounting footnotes
Following the adjusted-price thread, the broader version: any corporate action is a moment where your historical series and reality diverge unless somethin…
2
Aug 2026
When two data sources disagree, which one is wrong?
Ran the same symbol through two feeds for a week and compared. They disagreed constantly, not on price, but on volume, on the number of updates, and on exa…
2
Aug 2026
Snapshot data is not tick data, and knowing which you have changes your conclusions
Some markets distribute intraday data as periodic snapshots rather than every individual trade, a picture every few seconds, aggregating whatever happened…
2
Aug 2026
What the opening auction data actually contains
Following up on my automation post, here's what I've worked out about the data itself, since I'd never used it.
2
Aug 2026
Multi-timeframe context in three steps
I used to have four charts open and no framework. Now it's three questions, always in this order:
2
Aug 2026
A bar hides more than it shows
Two sessions with the same open, high, low and close can be completely different markets. One trended smoothly, one whipsawed six times. Your bar chart say…
2
Aug 2026
I found a beautiful pattern and it was a holiday
Spent an evening excited about a recurring dip at a specific time. Very consistent. Genuinely looked tradeable.
3
Jul 2026
Most "research" here is looking at charts until something appears
Said with affection, including about myself.
7
Aug 2026
Recording your own data changes what questions you can ask
Once you keep your own tick history, you stop being limited to the questions your charting tool anticipated.
3
Jul 2026
How do you decide a research result is real?
I keep finding patterns and I don't trust myself to judge them.
5
Jul 2026
Scanners find movers, not opportunities!
Took me a while to internalize. A scanner tells you what has already moved. That's the definition, it's ranking by something that happened.
4
Jul 2026
Funding rates are data, not just a fee
I treated funding as a cost line for a long time. It's also information. It tells you which side is crowded and how much they're paying to stay there.
5
Jul 2026
Crypto data quality varies more than people expect
Same asset, different venue, genuinely different data. Not just price by a few basis points. Different tick frequency, different behavior during fast moves…
4
Jul 2026
What features actually make sense for intraday? Genuinely asking
I keep reading about feature engineering and every example is either trivially obvious or clearly overfit.
5
Jul 2026
The order book is a snapshot of intentions, and intentions change
People read the book as if it were a commitment. It isn't. Resting size can be pulled instantly and frequently is, especially the size that looks most reas…
2
Jul 2026
Daily data is enough for more strategies than people admit
There's a quiet assumption here that finer data is always better. For a lot of approaches it just adds noise and decisions.
3
Jul 2026
Sample size, in the only terms that matter
People ask "how much data do I need" and the answer depends entirely on how noisy the thing you're measuring is.
4
Jul 2026