Daily cards for NBA are finally here. Wanted to share a couple notes for the new follows and answer some frequently asked questions.
I run a simple NBA model that uses offensive efficiency (points per 100 possessions), defensive efficiency (points allowed per 100 possessions) and pace (possessions per game) to project scores for each game. I calculate all the input parameters myself (using the data available here: https://docs.google.com/spreadsheets/d/1ph3j0jWhn66SmIweumX4nt2DXJUbXG-qUE_5Go66hVk/edit?usp=sharing) and adjust them for strength of opponent. After coming up with a baseline projection, I adjust for situations like travel, injuries, key player rest, trending performance and familiarity of opponents to result in a projected game score which is then compared to the line and total to find an edge and determine plays.
Summary of 2016–2017
The regular season was up and down. Most nights put together a 5-play card with a mix of spreads and totals with an average juice of -110. Generally took off Sundays and had a little hiatus back in November but playing consistently throughout the season was a full on grind. In total after 604 plays, finished with 309 wins, 284 loses and 11 pushes for a net -3.4 units. Here’s a nice visual aid of the highs and the lows:
Peaked at +19.1u about a week before the All-Star Break but then went into a free-fall the Wednesday and Thursday pre-break. Didn’t snap the cold streak until post-St Patty’s Day, officially bottomed out at -10.6 units. Despite posting plays on 118 days, only got swept twice (once in December and once in February) and only managed to sweep three times.
In general, overs were the cash cow hitting at 59% of over plays while under plays were atrocious hitting only 47%. No significant bias on the favs or dogs with both coming in around 52%.
The playoffs went well finishing 83–83–0 for +18.9 units. Key to winning in the playoffs was selling points (avoiding juice) and playing alt spreads when you identify your edge.
A Few Expectations:
· The model will be run for every game and projected score will be posted on weekdays by 3:30 PM Pacific time.
· Saturdays will likely have plays but may not post a full card with projected scores unless I have time to make the necessary adjustments across the full slate (and time to make the graphics).
· Similar to last year, we’ll be taking Sundays off with a few exceptions.
· Looking for the strongest edge in any given game and rarely will play both a side and total in the same game.
· In evaluating the edge we’re looking for over 2.5 points to make a play but sticking with 1u flat betting system unless there is evidence that the magnitude of the edge correlates with a higher win %tage for NBA (which will be evaluated after we have a month or two of data).
· Looking to sell points on almost every play, reduce the juice is the name of the game in the NBA.
· Doing all my NBA at 5Dimes where lines go up early and you can sell half points before the markets get sharp.
· All plays will be posted and tracked on the SportsAction App (a mandatory mobile app for sports betting information); because it’s kind of a burden, there may some small differences in the lines and odds on the app but the odds posted on the twitter feed are all real wagers placed at 5D and represent my official plays.
· Weekly summaries will be posted on Mondays, tracking results here daily: https://docs.google.com/spreadsheets/d/1fNLaqceIohND5NNZqmlEXD6ka1Yog3sLGxi7r2b4pDc/edit?usp=sharing
· Happy to answer any questions, hit me up whenever.