MLB Today's Best Bets (August 20): Gerrit Cole at Camden Yards Headlines a Nine-Game Getaway Board

Gerrit Cole takes the ball at Oriole Park at Camden Yards on Thursday evening, the marquee name on a nine-game board that empties out early. The job of this column is to set Verdexed's model probabilities next to the market's prices and report the gap. For a run of days that has not been possible, and Thursday is no different: verdexed.com was unreachable at publication, and the per-game win probabilities, confidence tiers, and edges were not retrievable through search either. No edge is claimed below and no confidence tier is assigned. That is a data availability limit on this column, not a verdict on the model or the slate. What follows is the market's read alone, labeled as such throughout, published for information and entertainment only, not betting advice, and for a reader 21 or older located where sports wagering is legal.

How to read this

A normal edition carries two numbers per game: the model's probability, and the probability the market implies once the sportsbook's margin is stripped out. The gap is the edge. Verdexed's published methodology builds the first from features covering recent form, team strength, home-field advantage, rest, and, in baseball, the starting pitcher, using only information available before first pitch. A calibration curve fit on held-out games is applied, so a stated 65 percent is meant to be a real 65 percent, then anchored to the de-vigged market price. Every prediction is logged before first pitch and scored afterward on win accuracy and Brier score. Verdexed's stated position is modest: markets are efficient, a large edge over the closing line is rare, and a small honest edge beats a fabricated large one. Two limits apply today. Trailing 30-day accuracy and Brier figures were not retrievable, so no hit rate is cited below, and per-game probabilities were not retrievable, so nothing below is called an edge. A well-calibrated 60 percent side still loses 40 percent of the time, and that is what 60 percent means.

The model's top edges today

There are none to publish. An edge exists only when a model probability and a market probability are observed at the same moment, and only one was available Thursday. Rather than fill this section with a number not read off Verdexed's board, it stays empty. A thin edge day and an unreadable edge day are different findings, and Thursday is the second.

Highest-confidence sides

What follows is the market's ranking, not the model's: the firmest prices surveyed on Thursday's board, offered as the comparison point a model number would be measured against. Every starter below is a probable unless noted as confirmed, and prices vary by book.

**Angels at Astros, 8:10 p.m. ET, Daikin Park.** The largest price on the board. Houston was quoted near -188 with the Angels near +158, a gap built on a wide standings separation more than on Thursday's pitching. Peter Lambert (8-6, 3.09 ERA) has been Houston's most dependable starter this season, and he was listed against the Angels' Grayson Rodriguez, acquired from Baltimore in the Taylor Ward trade. Starter listings here conflicted across sources at publication, so treat the matchup as unconfirmed.

**Blue Jays at Rays, 1:10 p.m. ET, Tropicana Field.** The series finale and the second-shortest price surveyed, Tampa Bay near -172, Toronto near +144, total 7.5. The favorite's case does not come from ERA: Ian Seymour is 9-3 but carries a 4.11 ERA, and Shane Bieber (4-2, 4.99) has the better name and the worse number. A calibrated probability would be doing real work here.

**Nationals at Rangers, 8:05 p.m. ET, Globe Life Field.** Texas was near -133 with Washington near +112 and the total at 7.5. Kumar Rocker (4-9, 4.50 ERA) is listed for Texas and has had a rough August, reportedly 0-1 with an 8.10 ERA across three starts. Washington's starter was the least settled on the board: Cade Cavalli pitched Wednesday, and Andrew Alvarez was the reported projection, unconfirmed at publication.

**Mariners at Brewers, 2:10 p.m. ET, American Family Field.** Milwaukee entered at 78-48, the best record in the majors, and was quoted only near -126, with Seattle (59-67) near +108. The reason is the matchup: George Kirby (8-9, 3.89 ERA) is the better starter on paper against Robert Gasser (3-5, 4.54). A wide standings gap against a short price is exactly where a model number matters most, and it is the one this column cannot supply today.

**Yankees at Orioles, 6:35 p.m. ET, Camden Yards.** Close to a coin flip once the margin comes off: New York (70-55) near -118, Baltimore (61-65) around even money. Gerrit Cole (6-6, 3.22 ERA) starts against Kyle Bradish (7-11, 3.65), two pitchers whose ERAs read far better than their records. Both were confirmed for Thursday, which makes this the board's most stable input.

**Cardinals at Reds, 12:40 p.m. ET, Great American Ball Park.** The earliest game and one of the odder prices: St. Louis was a road favorite near -114 with Cincinnati near +106. Michael McGreevy (5-9, 3.48 ERA) is confirmed against Brady Singer (5-11, 4.66). A road favorite in a hitter's park is a price a calibrated probability would resolve quickly, and none was available.

What could go wrong

The starter list is again the fragile part. Two of the six games above rest on probables that conflicted across sources, and a scratch at batting practice changes the most important input a baseball model has. Getaway day adds a wrinkle: six of the nine games start before 3 p.m. ET at the end of a series, when regulars sit and bullpens arrive short, and neither shows up in a season-long record. Weather is the usual variable on totals. The largest risk is already named: with no Verdexed probabilities in hand, nothing above has been checked against the model's numbers, so a reader is holding the market's opinion and a set of pitcher names, not a measured edge.

Responsible play

**This is model output, not betting advice.** Verdexed publishes its model's probabilities and edges for informational and entertainment purposes only. Nothing here is a recommendation to place a wager, and no outcome is guaranteed, even the model's highest-confidence picks lose regularly.

**You must be 21+** (or the legal age in your jurisdiction) and located where sports wagering is legal to bet. Sports betting is not available everywhere; know and follow the laws where you are. Never bet more than you can afford to lose, and never chase losses.

If gambling is affecting you or someone you know, help is available. In the US, call or text **1-800-GAMBLER** (1-800-426-2537), or visit ncpgambling.org. Please play responsibly.