The Daily Lob

Gray and Fried are separated by 0.08 runs of ERA and 1.36 runs of expected ERA. Gray has been carried by 1.15 runs across 169 innings; Fried's 2.80 sits against a 2.51 xERA.

Fried's 2.4% barrel rate is the lowest of any pitcher this page has carded in a fortnight, and an 85.9 mph average exit velocity says it is real rather than a sample artefact. Boston are down 0-1 and need the flattered pitcher to hold.

Against it  Fried has thrown 112.2 innings to Gray's 169 and his last start was 6.2 innings and four earned. Gray walks 2.4 per nine and has not allowed more than three earned in any of his last three. Elimination games also compress bullpens: the starter may only need to be fifth-best for four innings.

Sánchez threw 207 innings with a 2.91 ERA; Mahle threw 149 with a 3.99. On expected ERA the gap halves, from 1.08 runs to 0.56, because Sánchez has been helped by 0.48 and Mahle by nothing at all — 3.99 against 3.95.

Even halved, that is the largest true starter edge of the four games, and the market prices Philadelphia at barely 52% in a game they must win.

Against it  Mahle's last outing was five innings and five earned; his two before that were excellent, including seven innings of one-run ball against these Phillies on 12 September. Atlanta lead the series and can lose today without going home.

Yesterday both listed starters were relievers. Today Burke and Brown are real starters with 172.1 and 107.1 innings, and they are close: Brown better by 0.15 on ERA and 0.46 on expected ERA.

Houston are 57.7% favourites while trailing the series, which is the market saying the Game 1 loss was noise. Brown's 86.2 mph average exit velocity is the second-best contact profile on the board.

Against it  Brown walks 4.6 per nine, the highest of the seven starters carded today, and has not completed six innings in any of his last three. Burke's floor is lower still: 1.1 innings and seven earned on 8 September is in this three-start window.

San Diego go with Pivetta on 34.2 innings for the season — a 2.86 ERA against a 3.20 xERA, with a 10.1% barrel rate that is the worst of anyone pitching today.

The Cubs are still listed TBD after being shut out 8-0, which usually means a bullpen game or a short leash on whoever opens. Half this card cannot be written.

Against it  A 140-batter sample is not a season, and Pivetta's expected line rests on it. If Chicago open with a reliever the starter comparison is meaningless, exactly as it was in Houston yesterday — where the bullpen game won.

Twelve Elo points separate them and home ice is worth fifty, which makes Toronto a 58.8% side. The market has moved from a true pick'em at the open to 52.1%, still 6.7 points short of the rating.

Sorokin is listed as likely for the Islanders and Bobrovsky is unconfirmed for Toronto — which is the second night running that Toronto's goalie is a question the rating cannot see.

Against it  These remain preseason ratings with one game of evidence behind them. The market moving off the opener towards Toronto suggests it is finding the same thing slowly, and a line that is still moving is not a line to argue with confidently.

Colorado at 1570 against Los Angeles at 1501 makes them a 66.5% home side; the market says 61.8%.

Daily Faceoff projects Scott Wedgewood for Colorado against Darcy Kuemper. Neither is confirmed, and a team rating does not know which goalie is dressing.

Against it  Kuemper is the better goaltender of the two projected and that is a real part of why the market sits below the rating. Take the 4.7 points as a statement about skaters only.

Philadelphia opened at −125 and are now −141, which is most of a goal of market movement on a game Elo makes 58.7% to their 56.2%.

Eleven Elo points between the teams is nothing; this is a home-ice game and both ratings agree on that much.

Against it  A 2.5-point gap after a move that size is the market arriving where the rating already was. There is no edge left worth naming here.

Elo makes Arnaldi a 57.4% favourite on hard courts; the market makes Sakamoto one at 57.0%. That is 14.4 points of disagreement on a first-round match.

The ranking gap points the same way as Elo — Arnaldi is 37th, Sakamoto 137th — so the market is pricing something neither number contains.

Against it  Almost certainly a rising player Elo has not caught up with, plus a home crowd in Tokyo. Elo is backward-looking by construction and a 20-year-old climbing fast is exactly where it lags worst. A gap this wide against a liquid market usually means the model is missing information, not that the price is wrong.

Djokovic's hard-court Elo of 1997 is fourth in the world and 224 points clear of Borges, which translates to 78.4%. The market has him at 68.0%.

He is 39 and ranked 11th, and the market is discounting him accordingly. Elo, which weights results rather than age, is not.

Against it  The market has been right about ageing players more often than Elo has. A twelve-month rating window still contains a version of Djokovic that may no longer turn up for a second round in Beijing.

Tirante 1735.8 against Struff 1654.0 on hard courts gives Tirante 61.6%. The price says 61.8%.

There is nothing to say about this match that the number does not already say, which is worth one line rather than five.

Against it  Struff at 36 is capable of a serving week that no rating anticipates; that is a reason to leave it alone, not to bet it.