Hey guys im developing an algo to beat spreads in NBA games, so far its pretty mediocre. ive heard the refs officiating the games are a good thing to include inside a stats model, anything else? any sorts of key stats?
-
-
I'm doing the same. Mine does not take individual team stats into consideration it's more to identify good spot games.Last edited by myoung81; 12-09-16, 04:56 PM.Comment -
I'm using team record, past performance ATS and outright, home/away, underdog/favorite, days rest, OT, and same info for opponent right now but may expand.Comment -
NBA seems fairly situational. the games come fast and furious and it's quite a mental and physical grind for the players. that would be one of my key underlying theories.Comment -
You are wasting your time with anything like ats performance, individual team home/away performance, team records. Anything that you are able to find easily or that books provide on their stats pages will provide you with no help and will likely do you harm.
To build a good model, you need to be able to project team scores, taking into account injuries and minutes of players. Worry about minor things such as officials after you have a decent model that is being the close and profiting.Comment -
That, or just look for the right situations. As an earlier poster said, NBA is all about determining which situations produce abnormally high/low effort and energy in teams. I've realized a 1.6% edge over the last six seasons (700+ games) strictly using situational systems. Not saying you can't win with a model as well, but it's certainly not a requirement for success.Comment -
Days of the week and different months produce differing results, just like baseball. For an away team, I've used the amount of away games won in the past three away games as a starting point....teams that have won 1 or zero out of their past three away games tend to play better than the norm. The advantage is small and depends on the day, and I use some other filters to produce plays....sometimes though you will get a very small sample size which brings you the question....
"how small of a sample is too small?"
Angles/trends tend to be very fluid in the NBA...what works one year won't work the next commonly. This year home favorites with a worse winning percentage than their opponents have been very good in both the NBA and NHL.....but going back past this year, that trend hasn't stood up to scrutiny.
The NBA in my opinion is the most agonizing and difficult game to beat...there's a ton of bad beats in the last couple of minutes that are meaningless to the straight up result. Got beat a few week ago when I had a team +6 and my team, the dog got beat when they were down 5 with less than three seconds left and decided to foul.....stuff like that happens ALL the time. Every other sport has more predictable results and the effort of the teams and individual players is higher than the players/teams in the NBA.
I've waved the white flag numerous time in the past 30 years of NBA wagering....if you or anyone has consistent winning results in the NBA...you're a much better man than me.Comment -
I agree that the NBA is painful to bet (painful to watch as a fan also, IMO), but I do think there are situations that have held up over the years. This year, many of them have looked iffy, however. Still too early to tell if the market has adapted or it's just a bad run.Comment -
I am being disrespected by the same guy who asked if there are any sports one cannot "module correctly"? Ha. Somehow I'll get over it.
I'm sorry you didn't enjoy my two posts. Feel free to point me to anything public you have accomplished in the realm of handicapping.Last edited by evo34; 12-18-16, 10:12 PM.Comment -
that's one of the things i've found......
you backtest something that works really well and seems pretty reasonable i.e. not crazy data mining......... but then you look at current season and it's 34%............ so what do you do?
it's hard to find something that's amazing without huge risk of data mining.Comment -
Comment -
betting according to trends is like the people who go to a a roulette table and say "it was red 3 times in a row surely it will be black now"Comment -
@gojets It's a tough call. One thing I look at is avg. margin of victory for the system picks. If it's 12-20 ATS YTD and the avg. margin is not very negative (close to zero), I won't bail on it (assuming awesone historical track record, of course). But in general, if it looks like the edge might be decaying rapidly, I'll stop betting it and re-evaluate in the off-season. I think it's best to play it safe when in doubt.
For example, almost every decent NFL angle is failing in a huge way this season. I'm talking about ones that have worked fairly well for 10+ years. Variance? Change in situational peformance? Increased handicapper skill/awareness? Prob. all three. But the half-life for legitimitely good angles has decreased significantly over the past ~3-4 years. So a healthy dose of paranoia is always in order IMO.Last edited by evo34; 12-19-16, 12:25 AM.Comment -
The 1.6% ROI (presently 1.8%) is realized, not back-tested or theoretical. And that's without any line shopping. P.S. It's not a single system.
Record your bets, measure your performance. Not really a challenging concept, but I quite sure you've never done it.
Why don't you join PickMonitor and see how it goes?Comment -
haha man I'm loving this
I played like 500 bj hands a few weeks ago
came back with a 2% roi
My new system is killing it lately
trust me when I say this
If you can't calculate your edge on every bet you do then you are a losing bettorComment -
The fact that you play blackjack at all tells me you're not a serious handicapper. And clearly you have no concept of statistical significance.
Let me know when you join PM, and please predict your precise edge ahead of time. Can't wait to see all of your bet outcomes line up magically according to your expectations!Comment -
Or are you afraid to document your performance?Comment -
"If you can't calculate your edge on every bet you do then you are a losing bettor"
Only rank novices believe this fallacy. Have you ever even placed a sports wager? It certainly seems like you have zero experience outside of casino games, which is odd, given the attitude.
Anyway, I sincerely hope you figure out how to "module" every sport some day. :-)Comment -
Holy sh!t. I just realized you are this guy:
Quite possibly the dumbest post in the history of SBR. I cannot believe I wasted keystrokes even replying to your sorry ass. As I said in that thread, it's prob. time to retire this sub-forum.Comment -
I have never seen someone so thick in my life
There is an obvious line drawn between the troll and the serious post
You keep trying to discredit me by looking through my post history but what I have said in this thread is all facts
You seem to be someone who doesn't know what he is doing and flipped a coin a few times and think he is a winner
This is just a small snap of a few months of betting
Ignore the small amount since it was needed to find a phone number
I would never tell you what I bet because I am not stupid
The casino line was a play on words but you seem to be a bit too stupid to catch onComment -
i found something but of course it could be data mining we will see...Comment -
my nfl algo is going strong at least 60% each week.Comment
