troubleshooting Elio to retrieve the most up to date news around the NBA and broader sports world.
there’s currently a problem with the live market connector which is making some runs stale…
more to come shortly
we are very close to finalizing the product
8
Elio’s ML model is 2-1 so far on tonight’s NBA slate. We unfortunately missed a huge win on the Pacers, but we are still up units.
Elio seems to be really coming together…A lot of promise for real profit here going forward
Here’s a breakdown of plays on tonight’s slate thus far:
Sixers +170 ✅
Grizzlies +165 ✅
Pacers +300 ❌
Total +2.35 Units🚀
Elio is leaning Sixers ML tonight at +170 as well as the Pacers to beat Houston at ~+300
This will be a huge test. Does Elio love the long shot or is he finding a true edge in underdogs?
#machinelearning #sportsbetting #nbamodel #predictionmodel
24
early wins - underdog identifiers.
tonight Elio correctly located an edge and executed on Pistons +330 Moneyline.
Far from complete…more to come before our official launch
19
The hard part of a backtest is finding out what could break it. For Elio’s NBA moneyline research, that means checking what happens when you remove the biggest wins.
A model can pick winners and still overpay. Elio starts with two numbers: the model’s win probability and the probability implied by the price. The difference is where the research begins.
1
20
What would convince you an NBA model has an edge? A profitable backtest, consistently beating the closing price, or a forward test with every pick recorded before tip-off?
1
13
The Elio prediction engine and dashboard are starting to come together.
Working on turning all the modeling and backtesting into something clean, transparent, and actually useful.
Ran a bunch of NBA backtests today. Some promising results, some that completely fell apart once we dug deeper.
Honestly, finding out why a model fails has been just as valuable as finding what works.
1
16
One thing I’m learning building Elio is that predicting the winner and finding a profitable bet are two completely different problems.
That’s really where most of the work is going right now.
1
12
Been spending a lot of time figuring out which variables actually move the needle in NBA predictions.
Starting to see some interesting results from player-level data. Still a lot to test.
1
12
Originally started building Elio around NBA totals, but moneylines are looking more promising.
Interesting how quickly your direction changes once you start testing against actual sportsbook lines.
1
19
