I use the difference between my predicted score and the line. I align this value (I refer to it as a delta) up with a historical win rate from my backtesting... in most winning models, as the difference increases so does the win pct. I then account for simultaneous bets and go from there. I use estimates even for the kelly stake amount just to make life easy. Besides, once you hit the limits you're back to flat betting anyway... so kelly is over-emphasized. I don't know many people who are consistently better than 54% against closers when the limits are big. At that success rate, even kelly stakes are relatively small from a percentage standpoint, so just go 0.75 kelly at that point I guess. Unless you have a monster bankroll (>200k) splitting hairs over kelly stakes is kind of annoying.
-
Comment -
i don't know what else to say, other than go read anything you can find from Ganchrow and MonkeyF0cker. the link I posted previously encompasses this entire thread.Comment -
That's part of the noise. Certainly that situation would be an exception. You have to accept that those situations will occur and may cause additional losses... Over time that noise evens out. There are times when those types of changes work to your benefitComment -
did you call a closing line "noise" oh my heavens, LOL. tell me do you have an edge here
flip coin 1000 times
550 heads and 450 tails
that means heads would be a 55% winner on your kelly stake on the next heads? Think about that dude, wins and losses are not indicative of future performance in most cases and 99.99% of us fall into the most cases!Comment -
say a game closes at +100 and you won 60% on 1000 +100 bets what do you think your chance is of winning that +100 bet? and we are talking a sharp market like MLB or NBA or EPLComment -
You are asking for advice but respond in an obstinate way. Until you change your approach and attitude, you'll continue losingComment -
continute losing what? what reference to losing our we talking about? you did not answer the question no need to go to insults, tough questions sometimes make people respond with some other topic or anger, try and focus dude!Comment -
please if you can answer those couple questions so we get an understanding of your math model using win %Comment -
If a line moves due to injury at the close, it does not impact your overall performance because you are just as likely to benefit the next time. You are flipping a coin 96 times out of 100 with elite model. So you are just hoping for 4 true outcomes. Those kinds of moves are noise.Comment -
If a model shows consistent predictive value... Meaning that the win pct increases with increasing delta, and the line movement generally agrees with the model predictions then it can be considered adequately predictive, making the coin flip analogy inapplicable. There has to be a founded hypothesis first, then a test. If both make sense, and a good z score is there then it is likely to be valid. Since I did this, and had 6 consecutive winning seasons (actual win percentages match Backtesting), I consider it valid... For simplicity assume I am referring to point spread... All odds -110Comment -
so answer this because you used a bunch of big words not in my dictionary, please keep it simple, I am not that intelligent.
if 1000 NFL games are -4 -110 (which is high but whatever) how many games will they cover over 1000 games assuming the line is effecient which we know the nfl line is very very effecient?Comment -
wait give me both answers
how many times will they cover in your model and how many times with they cover not in your model?Comment -
i think this is where so many talented people get confused and start believing they can predict coin flips
the reality is if you bet at -4 and IT CLOSES AT -4 bet $100 bucks you got a 50/50 chance to win $91 period I dont care how you got there.Comment -
sorry I am typing on my phone and my grammar sucks anyway, if i say something so stupid you cant decipher please ask and I will rewrite it.Comment -
This is really not very complicated. If you have the Patriots at -3 and your model says -6, you bet -3 and you bet it large.
If Tom Brady and the second-string QB crack heads bending over to pick up a football and are out for the game and the line moves from -3 to +3, did you have a bad bet? You bet the good numbers and let things fall where they may.
People develop numbers in different ways and some are more effective than others. They bet based on their confidence in their numbers. You don't bet less on games that you believe have a bigger chance of winning if you believe you have a good model.
Those with the good inputs continue on and those with bad inputs go away.Comment -
derailer you never bet big its one game, one game does not make or break us its a grind or we lose!
if 1000 NFL games are -4 -110 (which is high but whatever) how many games will they cover over 1000 games assuming the line is effecient which we know the nfl line is very very effecient?
answer this derailer pleaseComment -
danshan, you are dealing with people that are very advanced in their models with a history of winning with a confidence in winning going forward.
You may believe that it is all luck but you have to address these issues with the assumption that you are dealing with a professional level gambler and not someone that just turned on a computer and decided to put something together. It is both presumptuous to assume otherwise and disrespectful to the people taking the time to answer your questions.
If you think professional level gamblers are not betting appropriately, you should provide some reasoning for your assumption or start a dissertation on bet sizes which I can assure you, you will get responses on.Comment -
I dont think pro players are making bad bets, I am asking you if a game closes at -4 -110 how many times will they cover out of 1000 plays? there is only one answer if the market is effecient has nothing to do with who bet it.Comment -
You are bumbling around claiming that people that believe that they have an edge do not have the edge they believe they have. It is incumbent on you to state your position and reasoning and not waste people's time telling them they are betting all wrong and throwing things against the wall until you can get someone to give you the information you want.Comment -
I highly doubt anyone is hitting at 54% against the spread UNLESS they are beating the line at that rate at least except for variance. Dont forget the line we assume is effecient.
long term you get about whatever the line closes at as a win rate if you play -4 at -110 and it closes -4 -135 now you are talking a winning 54% or higher even BUT -4 -110 bet closing at -4 -110 will be at 50% win rate more or lessComment -
the only edge to be had is in the line that is where the confusion is
a game closes at -4 -110 and they bet -4 -110 they will win 50% of their bets, you cant deviate from that UNLESS the market is not effecient and we know in big markets the market is effecient.
no matter who picks it if it closes -4 -110 and you bet -4 -110 it is a 50/50 bet only edge is if it closes different enough to cover the margin and that is edge, the only edge!Comment -
Why don't you start all over with a new thread making your position clear and why someone like MonkeyFOcker who is a professional gambler and is highly advanced in mathematics should not assume a certain edge. That is the kind of argument people love to destroy and read this sub-forum for. At the very least, you will get some of the responses you are looking for.Comment -
MOnkey might have an edge but his edge is not against the market effecieny his edge is against the line as I stated.
I will start a thread now and we can discuss this more sure, thanks for the adviceComment -
Thanks Derailer. Good responses all around.Comment -
new thread started please reply there guys, would love to have this discussionComment -
again Kelly is not flawed because the reality is that everyone should use it and bet $0 because they have zero edge and the other 1% know what they are doing, so the thing WE ALL MUST DO is know if we will have an edgeComment -
to me this might just be different defintions of the terms
market effeciency basically means to me the market closing line is effecient.
so if you bet 1 second before close and you get -4 -110 you have no edge I dont care what your model says or your previous win loss record.
an edge is getting the line at a better price than the closing line and to me is the only real edge in bettingComment -
and how to determine positive expected value? the only way I know to determine positive expected value is to bet a better line than the closing lineComment -
you cannot have a good bet if the market goes against it
if you bet the Bears -7 and they go to -5 your bet was bad on that particular game win or lose
think of it like this
you risk 100 to win 100 at -7 and now you risk 60 to win 100 at -7 with the new line change and you think you still have a good bet when you put 40 more dollars at risk for the same bet. This is a bad bet no matter how you sum it up, no matter what your model shows.Comment
