MLB Bankroll Management UK: Kelly, Flat Staking, CLV | FirstPitch

The £4,000 Lesson I Wish I Had Learned for Free

Eight years ago, I had a £5,000 bankroll for baseball betting. By the end of August that year, I had £900. I had not lost a system; I had broken every rule of staking I now teach. I had backed too many “must-win” plays at five and ten percent of my bankroll. I had doubled stakes after losses to chase. I had ignored the fact that even a positive-EV bet loses 40 to 50% of the time in the short run, and a streak of ten losses in twenty bets is mathematically routine.

Bankroll management is the single difference between betting being a hobby that costs you money and a hobby that pays. Picking winners matters less than punters think. Sizing them correctly matters far more.

The reason this is especially true in baseball is the variance profile. MLB plays 162 regular-season games per team. The favourites win at a long-run rate of 57.5% at average prices around –142.6. That sounds like a comfortable edge until you run the simulation: even a bettor with a genuine three percent edge over the market will see drawdowns of 25-30% of bankroll across normal variance, multiple times per season. If you cannot survive a 30% drawdown without changing your behaviour, you cannot bet baseball professionally.

This piece walks through the staking systems that work, the ones that look clever but break under pressure, and the recovery psychology that keeps you in the game when the variance turns ugly. There is a UK-specific layer to all of this — Gambling Commission rules around affordability and self-exclusion shape what a healthy bankroll relationship looks like for a British punter — and we will close on that. The sport is brutal. The maths is honest. The discipline is the only thing inside your control.

Why Baseball Punishes Loose Staking More Than Other Sports

Football has roughly 38 league matches per team. Tennis has its slams and tour events. Cricket runs Test, ODI and T20 calendars. Each of those sports has a relatively contained sample of high-leverage events per season. Baseball plays 162 games before the postseason. The sample size is huge, the per-game edges are small, and the variance grinds.

The data confirms what experienced punters feel in their bones. MLB underdogs win roughly four out of nine games on a long-running average — about 44%. That is high enough that any underdog system looks broken for two-month stretches and brilliant for two-month stretches, sometimes the same system in the same season. The bettor who cannot tolerate the rough patches will exit the system at the moment of maximum drawdown, which is also typically the moment of maximum future expected value.

The structural problem is that MLB betting markets are wide and active enough to tempt over-trading. There are fifteen markets per game, fifteen games per slate, four to six slates per week. A punter who finds a system that works on full-game moneylines might be tempted to extend the same logic to F5, run line, NRFI, props, and live derivatives — multiplying the action without multiplying the edge. Action is not edge. Action is the disease that kills bankrolls.

Baseball also rewards line shopping more than most sports because the prices on a single game can vary by 5 to 10 cents between books. Long-running data on this is striking. Across multi-year samples, fans betting through one book consistently miss out on closing-line value relative to those checking three or four books. Closing line value is the most reliable predictor of long-term profitability — beating the closing line by even a small margin per bet compounds into substantial expected returns over thousands of wagers. Without proper bankroll management, line shopping becomes irrelevant because you do not have a bankroll left long enough for the edge to compound.

The rule of thumb that holds across every system I have tested: if your typical stake is more than 2% of your current bankroll, you are over-betting. Variance will eventually catch you. The maths gets uglier the closer you push to 5% per bet — at that level, even a strong bettor faces single-digit-percent probabilities of going broke during normal variance.

Flat Staking: The One Percent Rule and Why It Survives Everything

The simplest staking system is also the most resilient. Flat staking means you stake the same percentage of your starting bankroll on every bet, regardless of perceived edge or confidence. The most common version is the 1% rule.

If your bankroll is £2,000, your unit is £20. Every wager is £20 — the moneyline you love, the run-line you marginally prefer, the F5 totals you slightly favour. Twenty pounds. Every time. The discipline is the system.

The reason flat staking survives is mathematical humility. It does not require you to know your true edge on each bet, which is fortunate because most punters consistently overestimate their edge. It does not require recalculating after wins and losses, which avoids the variance-induced compounding problems of percentage-of-current-bankroll systems. It survives drawdowns because the unit size stays small relative to the total bankroll even after a long losing streak.

The downside is that flat staking is mathematically suboptimal. If you genuinely have a 4% edge on one bet and a 1% edge on another, betting the same on both is leaving money on the table over the long run. The maths says you should bet more on the higher-edge play. The Kelly criterion is the formalisation of that insight, and we will get to it next.

The compromise version of flat staking is “1.5% on standard plays, 2% on top plays.” This adds a tiny dose of variance based on confidence without committing to full Kelly. I used this for years before moving to fractional Kelly, and many sharps still use it.

The biggest behavioural advantage of flat staking is that it removes the temptation to chase. After a losing day, the unit size has not changed. The next bet is still 1% of the starting bankroll. Compare this to systems that recommend “betting bigger to recover” — those systems wipe bankrolls out at frightening speed. The Martingale double-up is the most extreme example, but any system that sizes up after losses has the same fundamental problem. Flat staking immunises you against that mistake.

For a UK punter starting out, flat staking at 1% of bankroll is the correct system for the first year of serious betting. It will not maximise your return. It will keep you in the game long enough to learn what your true edge is, which is the prerequisite for any more sophisticated approach.

The Kelly Criterion Applied to Baseball

The first time someone explained Kelly to me, the maths sounded like an unanswerable argument for betting more aggressively. The maths is correct. The conclusion is wrong, and learning why took me about three years and several painful drawdowns.

Kelly is a formula. It tells you what fraction of your bankroll to stake on a bet given two inputs: the price the book offers and your honest estimate of the true probability that the bet wins. The output is the stake size that maximises long-term bankroll growth.

The formula in its simplest form is f = (bp – q) / b, where f is the fraction of bankroll to stake, b is the decimal odds minus one, p is your estimated probability of winning, and q is one minus p. If a bet is priced at 2.00 decimal (even money) and you estimate the true probability at 55%, Kelly says stake (1 * 0.55 – 0.45) / 1 = 0.10, or 10% of bankroll.

Ten percent. On one bet. That number should make any sensible punter flinch, and the flinch is the right reaction. Full Kelly is the theoretical optimum if your probability estimate is exactly correct. In practice, no human punter knows their true edge to within a single percentage point. If you think you have a 55% probability and the truth is 51%, full Kelly is over-staking by an enormous factor and will eventually ruin you.

The second problem with full Kelly is variance. Even if your probability estimate is correct on average, individual bets carry massive variance. A full-Kelly bettor with a genuine edge will see 30% to 50% bankroll drawdowns multiple times per year as a routine matter. Most punters cannot tolerate those drawdowns psychologically. They reduce stakes during the drawdown, miss the recovery, and exit the system.

For baseball specifically, Kelly is most useful as a sizing comparison rather than an absolute stake. If Kelly says bet 4% on game A and 1% on game B, the four-to-one ratio is the useful information. The actual stakes you place should be a fraction of those numbers. Sharp punters typically use one-quarter Kelly or one-half Kelly — multiplying the Kelly output by 0.25 or 0.5 — for exactly this reason.

One worked example will make this concrete. Bet on the Dodgers moneyline at 1.71 decimal (–141 American). Your model says the true probability is 62%. Plugging in: f = (0.71 * 0.62 – 0.38) / 0.71 = 0.0837, or 8.37% of bankroll at full Kelly. At quarter Kelly, that becomes 2.1%. At half Kelly, 4.2%. For most punters, even quarter Kelly is more aggressive than is comfortable, and capping any single bet at 2% of bankroll regardless of Kelly output is sensible.

Fractional Kelly and the Drawdown Problem

Quarter Kelly is the staking system I have settled on after eight years of experimenting with everything else. The reasons are mathematical and psychological, in roughly equal measure.

Mathematically, quarter Kelly trades a small amount of long-term growth for a substantial reduction in variance. Full Kelly maximises growth but with brutal drawdowns. Half Kelly cuts the growth rate by about a quarter and reduces drawdown variance by roughly half. Quarter Kelly cuts the growth rate by about half and reduces drawdown variance by roughly three-quarters. The trade is favourable for any punter whose probability estimates are not perfect — which is every human punter.

Psychologically, quarter Kelly produces drawdowns small enough that you do not abandon the system during bad streaks. A 10% drawdown is unpleasant but recoverable. A 35% drawdown breaks most punters’ confidence. The system that survives is the system that gets to compound your edge across thousands of bets over years.

The other practical advantage of fractional Kelly is built-in protection against probability misestimation. If you genuinely have a 4% edge but you mistakenly think you have an 8% edge, full Kelly stakes you at twice the optimal level — a path to ruin. Quarter Kelly stakes you at half of what you thought was optimal, which is still over-stake but far less catastrophically.

The drawdown problem in baseball deserves its own paragraph. Even with a genuine 5% long-term ROI, a quarter-Kelly bettor placing 1,000 wagers per season will see a drawdown of 15-20% of bankroll at some point during the season as a near-certainty. The punter who treats that drawdown as evidence the system is broken will exit at the worst possible time. The punter who treats it as expected variance will hold position, continue executing the system, and recover. The mental work is harder than the maths.

One reminder. Kelly assumes you have an honest probability estimate. If your “edge” is really just betting your favourite team or following a tipster you trust without verification, no fractional adjustment will save you. Kelly amplifies whatever your true edge is. If the true edge is negative, Kelly accelerates the loss.

Sizing by Confidence Tiers Without Pretending to Know Too Much

The middle ground between flat staking and Kelly is a tiered confidence system. You define three or four tiers — small, standard, strong, premium — with predefined stake sizes for each, and you assign every bet to a tier before placing it.

The version I use is three tiers. Standard plays at 1% of bankroll, strong plays at 1.5%, premium plays at 2%. Premium is rare — three or four bets a week at most. Strong might be ten plays a week. Standard is everything else that meets my criteria for placing a wager at all. The cap at 2% is non-negotiable; no bet, regardless of how confident I feel, gets sized above 2% of current bankroll.

The system works for two reasons. First, it forces you to declare your confidence in writing before the result is known, which is a useful check against retroactive rationalisation. Second, it caps the damage that overconfidence can do — even if you somehow misclassify every bet as premium, the worst case is 2% per bet, which is survivable.

The hard discipline is staying honest about tier assignment. If you find that 80% of your bets are landing in the premium tier, you are not finding more edge — you are diluting the meaning of the tier. The premium designation should be reserved for situations where multiple independent inputs point in the same direction with high conviction. If you cannot articulate why a bet is premium versus strong, it is strong.

Tiered confidence systems are not as mathematically optimal as proper fractional Kelly. They do not adjust to the actual price you are getting on a particular bet. A “premium” play at +110 and a “premium” play at +260 are sized the same in this system, even though Kelly would size them very differently. The sacrifice in optimality buys you simplicity, which for most punters is a reasonable trade.

Calculating ROI and Closing Line Value

Most punters track wins and losses. Few track the metrics that actually predict whether their betting will be profitable over time. Those metrics are return on investment and closing line value, in roughly that order of importance.

ROI is the simpler one. It is the percentage return you have made on the total amount you have staked. If you have placed 500 bets at an average stake of £20, you have staked £10,000 in total. If your net profit is £400, your ROI is 4%. The number tells you how efficiently your bankroll is being deployed.

The honest truth about ROI in baseball is that 5% ROI over a large sample is excellent and 8% is exceptional. Anyone advertising 15% ROI on baseball betting is either running a small sample, cherry-picking, or selling something. The reason the ceiling is low is that the markets are competitive. Sharp money corrects mispricings quickly. Edges shrink as books get better.

CLV — closing line value — is more important than ROI for predictive purposes. CLV measures the difference between the price you took on a bet and the price the same bet closed at. If you backed a moneyline at 2.10 and it closed at 1.95, you beat the closing line. The implied probability moved from 47.6% to 51.3%, meaning the market eventually agreed your side was a 3.7% better proposition than you originally got.

The reason CLV matters is that closing lines are the most efficient prices the market produces. Beating them consistently means you are reading information faster than the market — finding the right bet before the price corrects. Long-run profitability tracks CLV almost perfectly. ROI can be noisy in the short term because of variance; CLV is much less noisy because it is essentially a measurement of price movement.

The practical implementation: log every bet you place with the stake, the price you got, and the price the bet closed at. Calculate the percentage difference. Average it across all your bets. If your average CLV is positive — even 0.5% — you have a genuine edge that will eventually translate into ROI given enough sample. If your average CLV is negative, no amount of variance will save you. You are buying assets at retail and the market is selling them at wholesale.

One sample-size warning. CLV needs hundreds of bets to be meaningful. A 50-bet sample with positive CLV could easily flip to negative on the next 50. Most punters who calculate CLV for the first time over a small sample either dismiss it as noise or get falsely encouraged by an early positive run. Both reactions are wrong. Track it consistently for at least a season before drawing strong conclusions.

The Spreadsheet That Pays for Itself

Every serious punter I know runs a spreadsheet. Not a betting app. Not a notebook. A spreadsheet, with columns and formulas, that tracks every wager from placement to settlement. The reason is simple: nothing else gives you the data you need to evaluate your own performance honestly.

The columns I keep are date, sport (in case you bet other sports too), market type, teams, stake, price taken, closing price, result, profit-loss, running bankroll, ROI to date, CLV to date, and a free-text “thesis” column where I write down the reason for the bet before the result is known. That thesis column is the most useful column on the sheet. Reading back through your own pre-bet reasoning, especially on losing bets, exposes patterns of flawed thinking that you would not catch from results alone.

The most painful but most valuable insight you get from a tracking spreadsheet is the breakdown of profitability by market type. Most punters discover, when they actually look at the data, that they are profitable on one or two market types and unprofitable on three or four others. The combined ROI looks acceptable, but it is being driven by a narrow subset of bets while the rest is bleeding money. The fix is obvious: stop betting the markets where you have no edge. Easy to recommend, hard to do, because most punters are emotionally attached to “their” markets.

The other useful slice is profitability by day of week, by team, by starting pitcher type. Some patterns will be statistical noise. Some will be genuine edges or genuine leaks. The only way to tell is sample size, which is the only way to tell anything in this business.

For UK punters, one technical detail. If you bet across multiple licensed books to find the best price, your spreadsheet needs a “book” column too. You will likely find that one or two books consistently offer better prices on certain market types — perhaps Bet365 on totals, William Hill on moneylines — and biasing your action towards those books for those markets adds CLV without changing anything else about your process.

The spreadsheet does not need to be elaborate. The discipline of filling it in for every bet, immediately, and reviewing it weekly is what produces value. The fanciest tracking system in the world is useless if you stop logging bets two weeks in.

Recovering From a Downswing Without Making It Worse

The first time I lost ten bets in a row, I did the worst thing I could have done. I doubled my stake on bet eleven to “make it back.” Bet eleven lost. I doubled again on bet twelve. Bet twelve lost. By the end of the streak my bankroll was down 35% on the month and I was placing wagers I had not properly researched, on games I would never have looked at in calm conditions.

That sequence is so common in baseball betting that experienced punters have a name for it: tilt. The drawdown is normal — variance produces ten-bet losing streaks for any system, profitable or not. The tilt response to the drawdown is what does the real damage. The bets you make during a tilt are not the bets your system would have made calmly. They are larger, riskier, and selected from a smaller and less rigorous mental pool.

The recovery rule I use now is non-negotiable: stake size after a losing streak is the same percentage of current bankroll as before the streak. If my current bankroll is now 80% of where it started, my unit size adjusts down proportionally. I do not stake the same absolute amount in pounds, because that would be sizing up as a percentage. I do not stake more to “recover” — that is the route to ruin. I shrink stakes with the bankroll and play the system out.

The second recovery rule is volume reduction. After a meaningful drawdown — anything in the 15% to 20% range — I cut my number of bets per slate by about a third. The reason is not that I have lost confidence in the system. The reason is that drawdowns often happen during periods when something has shifted in the market or in my reads, and reducing volume gives me more time to evaluate each bet carefully. After a couple of weeks of reduced volume, I usually return to normal cadence.

The third rule is the time-out rule. If a drawdown has produced strong emotional reactions — frustration, anger, urgency — I take 48 hours off betting completely. No wagers. No price-checking. No analysis. The reset is not about recovering bankroll. It is about recovering judgment, which is the prerequisite for everything else.

One additional comfort. Across long-term data, professional punters experience drawdowns of 25-40% multiple times per year. Those numbers come from punters with proven edges over multi-year samples. The drawdowns are part of the job. Learning to absorb them without changing behaviour is what separates working punters from the rest.

UK Responsible Gambling Tools That Belong in Your Workflow

Bankroll management is the financial side of responsible betting. The behavioural side is what the UK regulatory framework is built around, and any serious punter benefits from using the tools the framework provides.

The headline number worth knowing is that 2.5% of the UK adult population scores at the problem-gambling end of the standard screening tool (PGSI 8 or higher). That is not a statistic about other people — it is a base rate that affects everyone who bets, including punters who consider themselves disciplined. The gap between profitable disciplined betting and harmful pattern is narrower than most punters want to believe, and the UK system gives you tools to keep yourself on the right side of it.

The most useful tool, in my experience, is the deposit limit. Every UK-licensed sportsbook lets you set daily, weekly and monthly deposit caps. Setting these at amounts your bankroll plan calls for, and then forgetting about them, removes the possibility of impulsive over-funding during emotional periods. Increasing a deposit limit requires a 24-hour cooling-off period under UK rules, which is exactly the friction you want when tilt is talking.

The UK sits at the centre of a £15 billion regulated online gambling market, with around 22.5 million adults engaged with regulated gambling on a regular basis, according to figures from the UK Gambling Commission’s leadership. The point of citing the scale is not to celebrate it. The point is to underline that regulation exists because the market is enormous, and that every UK-licensed book is required to offer the responsible-gambling tools you should be using.

The other tools worth knowing: time-out periods (24 hours to 6 weeks), reality checks (pop-up notifications during sessions), session-length caps, and full self-exclusion through the UK-wide GamStop scheme. None of these are “for problem gamblers” — they are tools any sensible punter uses to enforce the discipline their staking system already requires. A 24-hour time-out after a difficult day is the same idea as the 48-hour reset I described in the recovery section.

The full landscape of UK regulation around MLB betting — licensing, KYC, affordability checks, dispute resolution — is the subject of how Gambling Commission rules shape your wagers, which picks up where this section ends.

Created by the "tipsbettingb" editorial team.