Keeping records ROI can't fake

Most bettors track profit and nothing else, which is the one number that takes longest to mean anything.

Most bettors track profit and nothing else, which is the one number that takes longest to mean anything. A record with a few more columns in it answers questions about your betting years earlier.

What to record

Every bet, seven fields
Field	Why it earns its place
Date and time placed	Separates early-week bets from late ones, which behave differently
Market and league	Edges do not transfer between them; a blended number hides both
Price taken	Half of closing line value
Closing price	The other half — the single most valuable column
Stake, in units	Makes months comparable as the bankroll moves
Book	Reveals which book is actually giving you the prices
Result	Won, lost, pushed or voided — the last two are not losses

The closing price is the column people skip and the one that pays for the whole exercise. Without it a record can only tell you what happened; with it, the record tells you whether you were right.

Why ROI alone misleads

Return on investment is profit divided by total staked. It is the right measure over thousands of bets and a badly distorted one over dozens, because a single longshot dominates a small sample.

Two identical seasons, two very different numbers

Both bettors place 200 bets at one unit each.
Both go 96-104, and both have the same process.
One of them also hit a single +900 longshot.
That one bet is worth 9 units, on its own.

One shows +5% ROI, the other -4%. The gap is one bet, and it says nothing about which of them is better.

This is also why an ROI figure with no bet count attached is not information. "+14% ROI" over 60 bets and over 6,000 bets are different claims entirely.

The three numbers worth watching

Average CLV. Available immediately, on every bet, and the earliest honest signal you have.
Profit in units, by market. NFL sides and NBA props are different businesses; averaging them hides which one is working.
Maximum drawdown. Not a measure of skill, but the measure of whether your bet size is survivable.

Keeping it honest

The failure mode is not laziness, it is editing

Bets left out of a record are almost never random. They are the embarrassing ones — the impulse bet, the chase, the one placed after a drink. Leaving them out converts a measurement into a story, and the story is always kinder than the truth.

Log the bet when you place it, not when it settles. Later logging is where selective memory gets in.
Include the promotional and free-bet wagers, marked as such, so they neither flatter nor distort the rest.
Never grade against the softest close you can find. Use the same reference book every time.

Log a bet with its price, closing number and stake — the tracker does the rest. Start tracking.