SMT 100 · First Principles

There are three places to trade a sports market: a sportsbook, a prediction market and a pick'em app. They differ in who is on the other side, who sets the price and how the price is shown.

A sportsbook posts a price on both sides of a game and takes the other side of your position. The wager slip is the whole deal on one card: game, market, selection, odds, risk, to win, payout.

One price can be written three ways: American, decimal and fractional. Every one of them hides the same fact, the win rate the position needs to break even.

Add up both sides of a sportsbook price and you get more than 100%. The extra is the book's fee. Strip it out and what is left is the fair chance of each side.

Every game carries three core markets. The spread asks about the margin, the total asks about the combined score, the moneyline asks who wins. Each one has a line, a price, or both, and each can move.

Player props and parlays are where the book's fee gets bigger. A prop often carries more margin than a spread. A parlay stacks the fee, leg on leg.

A prediction market trades yes-or-no contracts that pay $1 if right and nothing if wrong. The price in cents is the market's chance, with a cent or two of trading cost on top.

On an exchange you do not have to take the posted price. You can post your own with a limit order and wait. And you do not have to hold to settlement. You can sell the contract to someone else first.

A pick'em entry is two or more player selections, More or Less, that pay a fixed multiple if every selection hits. The multiplier is the price, and the payout table is where the app keeps its edge.

Every venue caps how much you can put on, and every venue has its own way out before the result. A book caps the stake and sometimes offers a cash out. An exchange caps you at what the order book holds and lets you sell. A pick'em app caps the entry and usually has no exit.

The three venues answer to three different sets of rules. Sportsbooks are licensed state by state. Prediction markets answer to the CFTC. Pick'em apps fall under each state's fantasy law.

Every venue checks who you are, how old you are and where you are before it takes a deposit. Once the accounts are open, one ledger across every venue is the record that grades everything you do.

A model does not pick winners. It prices the game, so you can set your price beside the market's price and measure the gap.

The opening line is the first price a book posts on a game. It is set on thin information and low limits, which is why it moves easily.

A price changes as information and wagers arrive. This lesson separates a line move from a price move, a drift from a sharp move, and reads the canonical game from the open to now.

The closing line is the last price before the game starts. It carries every fact and every wager that arrived after the open, which makes it the market's most accurate price and the reference a model is measured against.

Both sides of a market carry the book's fee. This lesson shows how to measure it, how to take it out to see the fair chance underneath, and which of the two figures a model compares against.

Books disagree. The same position can be posted at three lines and three odds at the same minute. This lesson finds the price that pays the most, and shows what the difference is worth over many positions.

Team strength is how good a team is today, with the players who will play. Measured stats become ratings, ratings become a projected score, and who is out moves the rating before anything else does.

The home team gets an edge. It is small, it differs by sport, and it has been shrinking. This lesson measures it in points and applies it to the neutral-field margin.

Weather, rest and travel change what a game looks like on the day. This lesson values each in points, says which market each one moves, and applies them to the canonical game.

Ratings say how good two teams are. The matchup says how one team's plan meets the other's. This lesson measures a matchup from the play data, turns it into points, and completes the model price.

Technical gives the projection. Fundamental adjusts it. That is your model price. Put it beside the market's price, measure the gap, and at kickoff check it against the closing line.

Before a model prices a live game, it is tested on games it never saw. This lesson sets up a holdout test, reads the result in dollars and as a win rate, and names the three ways a backtest lies.

The close is the market's most accurate price. A model is graded by whether the close moves toward its price, counted over many games. That count arrives faster than a win rate and says more.

When the model says 60%, it should win about 60%. A model that passed its backtest can drift as the league changes. This lesson sets up the check that runs every week and reads its result.

One bucket is a check. Every bucket, side by side, is the calibration table. This lesson builds it from 100 past forecasts and adds the swing that separates a flag from a verdict.

News is what everyone knows. Information is what the price does not hold yet. Research is the work of telling them apart before you act.

Four things move a line: injuries, lineups, weather, and money from traders the book respects. Each one leaves a different mark on the price.

A move from open to now carries three facts: how far, how fast, and across how many books. Read all three before you decide who moved it and why.

Team strength is how good a team is today, with the players who actually play. Confirm the list before you trust the price.

Weather, travel, rest, altitude and surface are the conditions around a game. Each one is measured, valued in points, and matters in some sports and not others.

A matchup is where one team's strength runs into the other team's weakness. It is measured from the play data, turned into points, and added to the projection like any other adjustment.

Every read ends with one test. Hold your price up against the line. If the market already has what you have, there is nothing to trade.

Before you take a position, write the best case against it. If the case against is stronger than your read, the market is telling you something.

Official reports first, then the people who cover the team every day, then nothing else. Confirm before you act, and write down when the fact was released.

The research note is four lines written before the start: the time, the price, the reason, and what would change your mind. It is the artifact every lesson in this Part has been building toward.

The whole Part on one game. Lessons 1 to 9 run in order on a single sheet, the sheet ends in the four-line note from Lesson 10, and the note is held against the Brief on the Board line for line.

A system is a set of sizing and price rules written before the game. This lesson sets the capital those rules act on and shows why a rule written in advance beats a decision made while the game is on.

The unit is the standard measure of position size. This lesson sets it at 0.4% of capital, builds the unit table from half a unit to five units, and shows the choice between a proportional and a fixed unit.

Many sizing approaches exist. The Kelly criterion is one recognized approach: it sizes a position from a win probability and the price. This lesson works it in dollars, one operation per row, and shows why WAGERBIRD® does not size by it.

Sizing sets one position. Exposure sets how much is at risk at once. This lesson gives the standard protocol in percent of capital and in units: 2% per position, 3% per game, 10% per day, and counts a team total and a game total on the same side as one exposure.

A system names the sports and markets it trades and passes on the rest. This lesson sets the scope, runs every candidate through four gates, and records a pass as a position with zero units.

Execution is two readings and one comparison. The Board states Units; the limit is your worst price. This lesson reads the units off the Board, turns them into a stake, writes the limit before the game, and decides fill or pass on one position.

The limit price is the worst price a position is taken at, as a line and odds together. This lesson writes it next to every position before placing, shops it across operators, splits a fill across stake limits, and lets the position go when the line moves past it.

Every rule from Lessons 1 to 7 fits on one page. This lesson lays out the rulebook line by line, fills it on $10,000, and tests one position against it.

A real edge still loses in runs. This lesson gives the probability of a run of losses, shows what sound sizing does to the drawdown a run produces, derives the longest run to expect in 200 positions at 53%, and sets the pause rule.

The review is the one occasion on which a rulebook line changes. This lesson sets the review schedule, lists what the review reads from the record, separates what it may change from what it checks, and works one review on 120 units risked.

Memory keeps the wins and drops the losses. A written record keeps both. The record is the only grade a trader can trust.

One row per position. Seven columns per row. This lesson defines each column and fills the first row.

The price on the slip is true for one minute. Record it and the minute together, at that minute. Anything written later is a guess.

The close is the market's last and most accurate price. Compare every row's price taken with the same venue's close, and count the rows that beat it, tied it and lost to it.

A win at −160 and a win at +170 are not the same win. Grade every row in units won or lost, from its size and its odds, not in wins and losses.

Add the grades, add the sizes, divide. The running return is net units over units risked, kept current after every settled row.

A record with gaps grades nothing. No row is deleted, edited or voided by the trader. The only mark added after the fact is a void the venue made.

One return hides where it came from. Cut the record by sport, market, venue, price range and size, and the same four summary lines show where the record is strong and where it is weak.

A coin with no edge loses five in a row about three or four times in 120 flips. A return on a few rows is noise with a decimal point. This lesson sets how many rows a figure needs before it is read.

The record is read at a scheduled review, in a fixed order, and three decisions come out of it: keep, cut, change. Then the Board row, column by column, against the record.