I started tracking the closing line on every bet in April 2024 because I was tired of not knowing whether my process was actually working. Twelve months later I have 1,483 bets logged, every one tagged with the closing price at the same bookmaker. The aggregate number is +2.7% average CLV. The breakdown by category is where it gets interesting.
The overall number
+2.7% is legitimate but unspectacular. It sits squarely in the "real edge, not elite" bucket from the CLV guide. It's the number I'd expect from someone who's doing the work, finding the obvious soft spots in the AU market, and not making any structural mistakes. It is not the number of someone with proprietary information or a unique edge. It's the number of someone who is paying attention.
Over 1,483 bets at average stake around $120, that CLV translated to about $4,800 of expected profit, and actual realised profit of $5,340. The gap between expected and realised is variance being kind to me over the sample. Both numbers are small relative to the stakes involved, which is worth internalising - advantage betting at retail scale is not get-rich money.
By bet category
The interesting part is where the +2.7% came from, and where it got dragged down.
Player props: +6.2% average CLV across 412 bets. This was most of the edge. AFL disposal markets, NBA points and assists lines, NBL player points - every one of these is priced by a smaller trading team than the H2H markets, and the cross-bookmaker dispersion is larger. Easier to find a +5% line on a disposal total than on an AFL H2H, and the closing price tends to tighten toward the median. If I had to recommend one area to a new AU advantage bettor, it's props.
Promo-boosted bets: +11.4% average CLV across 94 bets. This is artificially inflated. The boost itself is what makes the CLV look good - the underlying pick is usually only mildly +EV against market consensus. Still, a real edge, just one you can't scale. Every boost is capped at $50 or $100, and you can only hit each once per account.
H2H mainline: +1.3% average CLV across 487 bets. Much thinner. This is where the AU market is most efficient - lots of bookmakers quoting the same games, sharp money flowing in early, lines moving fast. Most of the bets here came from timing: catching early lines before they shortened. When I placed H2H bets in the final 30 minutes before kickoff, average CLV dropped to -0.2%. The edge was entirely in the placement time.
Totals and lines: -0.4% average CLV across 298 bets. This surprised me. I'd assumed totals markets were softer because they get less attention. What I was actually doing was betting totals on gut feel rather than against consensus, and gut was losing to the market. I stopped betting these around month nine and the overall CLV improved.
Multis and SGMs: -3.1% average CLV across 38 bets. I kept these in the data as a reminder not to touch them. Every multi I placed was negative CLV on average, which makes sense because multis compound the vig. The handful I won were pure variance wins.
Arbitrage legs: not included in CLV tracking because CLV on an arb leg is meaningless. You're not picking sides, you're locking in a hedge. These sat separately in the arb ledger.
By sport
The sport-level breakdown was less useful than I expected because the variance is high. Rough numbers:
- AFL: +3.1% over 414 bets
- NRL: +2.4% over 296 bets
- NBA: +4.9% over 338 bets (mostly props)
- NBL: +5.8% over 127 bets (props again, very soft market)
- A-League: +0.8% over 62 bets
- EPL: +1.2% over 84 bets
- NFL: -0.6% over 74 bets (I'm bad at NFL)
- Racing: not tracked in this dataset
Two takeaways. The Australian-origin markets (AFL, NRL, NBL, A-League) had strong edge at the player-prop level because local pricing teams are small. The international markets (NBA, EPL, NFL) had edge only where I was catching specific timing windows. NFL was straight-up negative, which was a clear signal to stop betting NFL and I ignored it for another four months because I liked the games.
By bookmaker
The slowest books gave the best CLV on early-market bets:
- BlueBet: +5.4% (gubbed day 63)
- BetRight: +4.8% (still open but heavily limited)
- Dabble: +4.3% (still open, modest limits)
- Unibet: +3.7% (limited)
- PointsBet: +3.1% (limited)
- Ladbrokes: +2.2% (limited)
- Neds: +2.1% (limited)
- TAB: +1.6% (still open, very modest limits)
- Sportsbet: +1.1% (limited)
- Bet365: +0.8% (limited, very fast book)
- Betfair Exchange: varies, generally +1.5% on backs
The pattern is exactly what you'd predict from the bookmaker business model. The soft books delivered the most edge and closed first. The sharp books (Sportsbet, Bet365) delivered less edge but stayed open longer. By month nine, half my CLV was coming from the three books most likely to cut me next, which is exactly the death spiral every arber and +EV bettor experiences.
The three weeks I almost quit
CLV protects you emotionally during losing streaks, but only if you trust the number. I hit a stretch in September where I went 17 bets without a single win over a three-week period. The average CLV over those 17 bets was +4.1% - legitimate edge, just a variance hole. Meanwhile the bank balance was down $1,400 and my girlfriend was asking careful questions about whether I should maybe stop.
The thing that got me through was looking at the CLV column in the spreadsheet rather than the P&L column. Every one of those losing bets was recorded with positive CLV. The process was sharp. The results were just unlucky. The next 30 bets hit at 57% win rate and I ended the quarter +$2,100.
If I'd been tracking wins and losses only, I would have changed something during that stretch, and the change would have probably cost me the recovery run. Nothing in betting is more valuable than a disciplined measurement system during a bad patch.
What I'd tell someone starting
Three lessons from the year.
Track from day one. You can't reconstruct closing lines retroactively. Every bet you don't tag with a closing price at placement time is a data point you'll never have. The Krok Odds Bet Tracker captures it automatically, which is why I built it - my own spreadsheet had a 30% fill rate and the numbers were unreliable for months.
Segment by category. Overall CLV hides what's working and what isn't. I was running +2.7% aggregate for months while losing money on totals and multis and not noticing because the props were carrying the whole operation. The moment I cut the leaking categories, the numbers tightened up.
Act on negative signals. NFL was negative-CLV from month three, and I kept betting NFL until month seven. That delay cost me something like $600. When the data tells you you're bad at a market, stop betting that market. This sounds obvious in print and is shockingly hard to do when you enjoy the sport.
The full year of data lives in my Krok Odds tracker and I look at it at least once a week. Twelve months in, the process is measurably sharp, the edge is real but modest, and the biggest remaining risk is account death. Which is a different problem entirely, covered in the gubbing guide.
David has been running advantage betting strategies across Australian bookmakers since 2023 and contributes long-form retrospectives, case studies, and operational pieces drawn from years of running real bets in AU markets. His writing focuses on the realities of running a sustainable AU advantage operation — what works, what fails, and the operational details most blogs gloss over.