Readers of my weekly column cashed in nicely with my first Double Bet of the 2018 MLS Season on Atlanta last Sunday.
The model was very accurate this week and appears to be coming into excellent form. The Minnesota Chicago and Philadelphia Columbus prices are close enough that there is no edge. The NY City Orlando Atlanta Vancouver and Salt Lake NY Red Bulls matches are within 0.02 of the current market price. The model suggests two bets on Toronto and Seattle, but it does not account for both teams travelling to Mexico midweek for CONCACAF Champions League matches.
That leaves two matches which stand out. Here is a full breakdown of the projections:
DC United vs Houston Dynamo
Minnesota United vs Chicago
Philadelphia vs Columbus
Montreal Impact vs Toronto FC
Atlanta United vs Vancouver
Sporting KC vs San Jose
Real Salt Lake vs NY Red Bulls
FC Dallas vs Seattle Sounders
DC United vs Houston is the first match of the season where the model has projected a draw price lower than market value.
As I write this, the best draw price in the market is 3.40. After I remove the bookmaker margin from the price, I can calculate a market price of 3.48, implying a 28.7% probability. These odds are well above my model price of 2.93 or a 34.15% probability.
What interests me more in this match, however, is the fact that the 0-0, 0-1, 1-0 and 1-1 scorelines (all under 2.5 goals) equal a projected probability of 59.63%. The current Under 2.5 market price is 1.97, a 49% implied probability after removing margins. This 10% difference is a couple of percentage points more than the advantage with the draw.
The league is playing well shy of its 22% average draw rate (just two draws in nineteen matches this season – and none last week). This match has all the projections for a low scoring draw. I will back a full stake on Under 2.5 and a half on the draw.
Sporting Kansas City is the most significant advantage identified by the model. The stand out data point appears next to their opponent, San Jose. The model has projected the visitors a whopping 65.8% chance of scoring zero goals and just a 27.5% chance to score one goal. The sum of those two probabilities (zero or one goal scored) equals 93.3%. In comparison, the model has a 60% expectancy that Kansas City scores two or more goals.
Not only is the home side projected to be a 60% chance to cover 93% of the visitors expected output, but Kansas City also comes out ahead in the total shot ratio 0.58 to 0.53 too.
I also like how the on-field narrative directly corresponds with the model predictions. Last season it took Kansas City – a side prided on defence – ten matches to concede five goals. This season they have allowed five goals in just two matches. A favourable home match means their focus will be on righting their strength (defence) which plays back into the percentages put out by the model. There is undoubtedly value in the current market price of 1.61 with Kansas City.
