Let us beat that old horse, staking method!

[danshan11]

Like everyone knows I dont like Kelly per se but I kinda use Kelly learned it from DRH on BT who got it from SBR I dont remember the author but anyway
I use
bet X.XX to win 1 on faves and .XX to win 1 on dogs
Reasoning
I like to have more money in on bets I am more likely to win. I know this method is probably a 3rd cousin of Kelly but I have backtested my last year NBA and NHL and I would have been up 10 more units for the season. 10 units to me is a huge huge move.

What methods do you guys recommend and what is flawed with my method? Dont pull punches I can take it I got headgear on.

[oilcountry99]

flatbet as per vanzack's true flat bet method which I learned across the street

[danshan11]

I dont really understand that method, can you in a few words tell me the advantages to that method?

[oilcountry99]

In a nutshell, it makes your risk/reward ratio equal across all bets. You will make more profits the closer the line is to even. It keeps your risk in check that is the biggest advantage in my opinion. anyone interested can google it and read the disertation

I'm not 100% sure but I think the profits would be better than the method you describe in post 1. The answer is probably in vanzacks thread.

[snapperman2]

I think that Stanford Wong recommended that you bet in such a way that your total return is always the same on each bet and that decreases your variance to a minimum. In other words, if you are betting at +100, you bet one unit for a total return of 2 units (your original bet + win). If you are betting at -300, bet 1.5 units for a total return of 2 units. If you are betting at +300 bet 0.5 units for a total return of 2 units.

[oilcountry99]

snapper...well said

[danshan11]

I think that Stanford Wong recommended that you bet in such a way that your total return is always the same on each bet and that decreases your variance to a minimum. In other words, if you are betting at +100, you bet one unit for a total return of 2 units (your original bet + win). If you are betting at -300, bet 1.5 units for a total return of 2 units. If you are betting at +300 bet 0.5 units for a total return of 2 units.

isnt that what I am doing?

[oilcountry99]

no, you are betting to win 1 unit no matter what the odds

[danshan11]

oh yeah ok duh, I see what is the advantage to that method? what do you think the disadvantage to my method is?

[danshan11]

I read that Vanzack method and I dont understand where the advantage is and how the bets are broke down

[snapperman2]

With your method if you bet a heavy favorite, you could be hurt badly if it loses. For example if you bet at -2000 you will lose 20 units on one bet. You are also not betting enough on dogs, and they are usually a better bet than favorites.

[danshan11]

I dont bet -2000 favorites I stick mainly to the bigger leagues but your point is very valid on favorites.

I really do not understand the betting too little on dogs, why do you think dogs are a better bet? Why do you think I am not betting enough on them.

doesnt it make sense to you to have more on things more likely to occur than the other way around?

[HeeeHAWWWW]

Betting to win a constant amount, rather than a risking a constant amount, has higher expected growth. That's why it's inherent to the Kelly formula.

[danshan11]

Betting to win a constant amount, rather than a risking a constant amount, has higher expected growth. That's why it's inherent to the Kelly formula.

does that mean my method is better than flat betting? and why do you think so?

[Derailer]

does that mean my method is better than flat betting? and why do you think so?

If you do not have CLV, do not bet at all!!!

[danshan11]

If you do not have CLV, do not bet at all!!!

amen very true! I only bet when I think I have an edge.

[Bsims]

What methods do you guys recommend and what is flawed with my method? Dont pull punches I can take it I got headgear on.

A few months ago I decided to look at a few bet sizing approaches to determine if there was a better approach for me based on the system I use to pick wagers. I started the thread "Kelly Re-Visited" in this forum. I also devised and executed a bet simulator that had an accurate edge. Finally, I documented this project in 11 posts on my blog (ole44bill.blogspot.com. They began on June 20th and ended on July 11th.

I think the most important thing I learned came from a post by KVB when he discussed the best approach. “It really does depend on your goals and time frame.” Based on this effort, I am using flat wagers for my largest wagers based on a long trusted approach where the odds don't vary much. This gives me the best chance of having a positive season. I'm using Kelly for some smaller systems I'm testing. This provides a better chance of hitting it big if one proves out and limits my losses if they don't.

I did not consider your approach or Stanford Wong's. Both interest me as a replacement for my flat bets. Note that both are essentially flat bets for sports like football and basketball where most odds are -105,-105. These techniques size the bets based on the odds and are not dependent on estimating an edge at the time of the wager. Kelly and similar are based on having confidence in estimating the edge.

If you feel you've identified a good wager but won't know your edge until the lines close, then the former is for you.

[danshan11]

very good reply Bsims, i think in the -105 -105 scenario it is even better to use my system because it rarely is exactly -105 -105 its usually -115 +100 or some variant and that makes my staking method priceless, it seems more times than not i get in a position (no factual back this up data) that I make 4 bets and say 2 are faves and 2 are dogs, my dogs lose and my faves win and I have a small +day because of it.
example
bet .9 to win 1
bet .92 to win 1
bet 1.1 to win 1
bet 1.1 to win 1
last two win I am up 2 units first 2 lose I am down 1.82 and I net .18 for the day

[ChuckyTheGoat]

I think that Stanford Wong recommended that you bet in such a way that your total return is always the same on each bet and that decreases your variance to a minimum. In other words, if you are betting at +100, you bet one unit for a total return of 2 units (your original bet + win). If you are betting at -300, bet 1.5 units for a total return of 2 units. If you are betting at +300 bet 0.5 units for a total return of 2 units.

Agree w/ Oil. This is a nice summary.

Riskier the play, smaller the bet-size. IE, lottery ticket merits no more than a few pennies.

Only thing I don't like is that this doesn't consider different plays have various ROIs. IE, you should also be scaling to higher-rated Hi-ROI plays (in some way).

[danshan11]

Agree w/ Oil. This is a nice summary.

Riskier the play, smaller the bet-size. IE, lottery ticket merits no more than a few pennies.

Only thing I don't like is that this doesn't consider different plays have various ROIs. IE, you should also be scaling to higher-rated Hi-ROI plays (in some way).

Have different ROIs what do you mean? how can a different bet have a different ROI?

[HeeeHAWWWW]

does that mean my method is better than flat betting? and why do you think so?

I'd guess so - although the Stanford Wong method would be better.

Imagine you have 100 bets, at a range of odds from -200 to +500. With fixed-risk betting, your end results are largely dependent on how a very small number of longshot bets perform. Win one or two, it's a drastic effect on your RoI.

Variance is your enemy, and betting to win a constant amount smooths that out.

[Miz]

Hehaww makes a good point. I have abandon working models that exhibited too much variance due to high payout bets making up most of the profit. The fluctuations suck in real time, but they are tolerable in a graph in a backtest. Psychologically and from a risk management standpoint, I prefer more of a 55% win rate with a -110 type of payout.

[Optional]

I have always thought blindly increasing your risk amount simply based on odds offered, without concern for edge, has to be a bad strategy.

[semibluff]

I'd guess so - although the Stanford Wong method would be better.

Imagine you have 100 bets, at a range of odds from -200 to +500. With fixed-risk betting, your end results are largely dependent on how a very small number of longshot bets perform. Win one or two, it's a drastic effect on your RoI.

Variance is your enemy, and betting to win a constant amount smooths that out.

You might consider splitting your bankroll into 2 pools. 1 for safer/normal bets and 1 for the long-shots that distort the overall results. That way the long-shots shouldn't influence staking size, roi, or edge analysis.

[danshan11]

I have always thought blindly increasing your risk amount simply based on odds offered, without concern for edge, has to be a bad strategy.

when you say edge, what do you mean by edge? my definition of edge is not known until the game line has closed, what do you mean?

[danshan11]

You might consider splitting your bankroll into 2 pools. 1 for safer/normal bets and 1 for the long-shots that distort the overall results. That way the long-shots shouldn't influence staking size, roi, or edge analysis.

I went back and removed all my super long shots from my bets my CLV stayed the same and my Units went way up

[tsty]

how many of you are actual pros that make in excess of 50k pa?

too much stupidity in thread

[tsty]

I have always thought blindly increasing your risk amount simply based on odds offered, without concern for edge, has to be a bad strategy.

you missed the part where betting in general without knowing your edge is stupid

from my recent readings it seems majority of posters cant even generate their own lines and are stuck in 2008 with this system nonsense

if you cant even create your own lines then this aint for you

[danshan11]

I surely know nothing and that is why I keep discussing it. I want to learn. I look at my history and I think it comes back to noise variance on the adjustments I would make to increase units. In reality I think any adjustment would just be treating symptoms not curing my disease. I like my staking method and will stick with it, I will learn the longshot fave bias so I can see what it is truly doing to my CLV net, I will continue to try and sharpen my lines so I have more betting options and increase the number of bets I make that are edge bets. I just am looking at 100 or 200 games and trying to determine something from a tiny sample size and that is flawed thinking.

[danshan11]

@ tsty what would your advice to me be from you reading what I have said on the forum in general?

[Miz]

how many of you are actual pros that make in excess of 50k pa?

too much stupidity in thread

who is going to admit on a forum they win large amounts ? lol

[Optional]

when you say edge, what do you mean by edge? my definition of edge is not known until the game line has closed, what do you mean?

you missed the part where betting in general without knowing your edge is stupid

from my recent readings it seems majority of posters cant even generate their own lines and are stuck in 2008 with this system nonsense

if you cant even create your own lines then this aint for you

Maybe you guys misunderstand me.

Whether you "know" your edge on the bet or not, surely increasing your risk amount blindly based on the odds, ie: betting 130 to win 100 or 100 to win 130 anytime the odds are -130 or +130, is not mathematically sound.

You guys seem to saying (assuming?) it is. Do you have numbers to support that?

[danshan11]

who is going to admit on a forum they win large amounts ? lol

I win large amounts of money on CFL NBA and WNBA and on the MLB I get destroyed. I am just now seeing if I have an edge on soccer(only because so many games so often)

[danshan11]

Maybe you guys misunderstand me.

Whether you "know" your edge on the bet or not, surely increasing your risk amount blindly based on the odds, ie: betting 130 to win 100 or 100 to win 130 anytime the odds are -130 or +130, is not mathematically sound.

You guys seem to saying (assuming?) it is. Do you have numbers to support that?

My big issue on betting my edge is I dont really know my edge before the closing line comes out
my reasoning of changinge bet based on the line is that I am assuming the line is fairly effecient, so changing bet size according to line is reasonable in that respect.

[Miz]

i encourage everyone to go back to threads in this forum from about 2008-2012... it will help you