Poker hand histories vanish at 60 days — your notes don't

Most Australian-facing poker rooms and tracking tools wipe or lock your hand history at the 60-day mark, and a handful cut you off at 30. Your own notes don't expire unless you let them. That gap — between what the platform keeps and what you keep — is the whole reason a plain text file can outperform a $200-a-year tracker subscription if you're playing anything below mid-stakes.

What actually happens at day 60

The 60-day figure isn't a coincidence or a conspiracy. It's storage economics dressed up as policy.

PokerStars, for instance, has historically archived hand histories after a set window and required you to request older data rather than pulling it from the client. Other rooms vary — some keep 90 days, some keep everything but only surface the last few thousand hands in the replayer. Third-party trackers add their own layer: many have free tiers that cap stored hands or limit how far back you can filter, and if you stop paying, the database doesn't grow.

Here's the practical problem. If you play 15,000 hands a month across two rooms, by day 61 you've got 30,000 hands of context sitting in a place you can't easily query. That's not a small sample. At a 5bb/100 win rate, 30,000 hands is enough to distinguish "running bad" from "actually losing" with reasonable confidence. Losing access to it means you're rebuilding reads from scratch every two months.

The rooms aren't doing this to spite you. Hand history storage at scale is genuinely expensive, and most players never look past their last session. The 60-day default serves the median user. You are probably not the median user if you're reading this.

The tracker trap

Paid trackers solve this by hoarding everything locally. That works — until it doesn't. Hard drive dies, you switch from Windows to Mac, the company gets acquired and shuts the servers down, or you just stop paying the subscription and the software locks you out of your own database. I've watched players lose four years of data this way. The hands were always theirs; the access was rented.

Why your notes survive where hands don't

A hand history is raw data. A note is a conclusion. The conclusion is worth more.

Think about what you actually use a hand history for. You're not re-reading 30,000 hands. You're answering questions like: does this villain check-raise the turn with draws, or only with made hands? Does he 3-bet light from the small blind? Does he fold to a river probe after calling two streets?

Those are patterns, not hands. And patterns fit in a sentence.

A note that says "SB vs BTN: 3-bets 22% over ~400 opportunities, but folds to 4-bet 80% — range is polarised, not merged" is worth more than the 400 hands it came from, because it's already been processed. The hands expire. The sentence doesn't.

This is also why notes travel better. A tracker database is tied to a specific room's hand format, a specific software version, a specific operating system. A text file with 200 villain notes works on every site you'll ever play, in every country, on every device, forever. You could paste it into a Google Doc and it would still be useful in 2035.

What a good note actually contains

Bad note: "calling station."

Good note: "Calls flop c-bets 70%+ on dry boards, folds turn to second barrel ~60%. Doesn't raise draws. Value bet thinner than normal, don't bluff rivers."

The second one is actionable in a hand you're playing right now. It has a frequency estimate, a board texture qualifier, and a direct instruction. It took maybe 20 seconds to write. It'll pay for itself the next time you're in a pot with that player.

The trick is writing them during play, not after. Post-session review is where good intentions go to die. If you're not noting the villain while the read is fresh, you're not noting them at all.

The 200-hand rule and why it's wrong

There's a folk rule in poker that you need 200 hands on a player before you can trust a stat. That's roughly where HUD numbers start stabilising for things like VPIP and PFR. It's also completely useless for note-taking, because by the time you've got 200 hands on someone, you've probably already played the important pots against them.

Here's a better frame. You need one specific observation to write a useful note. One showdown where a tight player showed down a hand you didn't expect. One river where a nit bluffed into three players. One 3-bet pot where a maniac flat-called with aces.

Single observations are noisy. But they're not worthless — they're hypotheses. Write them down. If the next observation contradicts it, update the note. Over 30 or 40 hands against the same villain, you'll have a sharper read than any HUD stat will give you, because you'll know why the numbers look the way they do.

The 200-hand rule assumes you're using aggregated stats. If you're using specific observations, the threshold is much lower.

The reg problem

Notes work brilliantly against recreational players, who are the majority of your profit. They work less well against regs, who adapt. If you've got a note saying "folds to 3-bets wide," a good reg will notice you exploiting it and adjust.

Two things fix this. First, date your notes. "Folds to 3-bets wide — March 2024" tells you the read might be stale. Second, note the reason, not just the behaviour. "Folds to 3-bets wide because he's opening too many suited connectors from EP" tells you what to watch for if he changes.

Regs also change rooms, change stakes, and change games. A note from two years ago at 50NL might be worthless at 200NL. Your notes age. That's fine — they age better than deleted hand histories.

What to actually do

If you want a system that outlives any platform:

Keep one file. Plain text, synced somewhere you control — Dropbox, a Git repo, wherever. Not a proprietary format. Not locked inside an app.

Write in it during play. One line per villain, screen name first, then the read. Update it when you see something new. Delete notes that are clearly wrong.

Review it before sessions. Not the whole thing — just filter for players you might see. If you're playing a 6-max pool of 300 regs, you'll see the same 40 names constantly. Those 40 notes are worth more than any database.

Export your hand histories before day 60. Every room lets you do this manually, even if it's tedious. Do it once a month. Store the raw files somewhere durable. You probably won't read them. That's fine — the point is that you can.

The uncomfortable question is whether any of this matters at stakes where the rake eats most of your edge. At 2NL and 5NL, note-taking is a rounding error against the rake. At 25NL and above, against a pool of regulars you'll see for years, it's the difference between a player who learns and a player who just plays.

Which raises the real question: if your reads are the only durable asset you have in this game, why are you letting a 60-day retention policy decide how many of them you keep?