Losing Steaks - I Need To Know

[cameron398]

Hello all. Considering all the methods and research completed in the past by members of this forum, I want to see some losing streak figures.

So members, despite your best efforts in making selections and betting methods what are some of the horror losing streaks you have experienced in testing your results ?

I am particularly interested in losing streaks where a member may have been betting on more than one horse in a race. However, any stats you have on file would be of interest.

In case I am not making my intentions clear I am trying to get a feel for what kind of losing streaks may be possible, even when selecting "in form" or "promising" horses.

In all my browsing I only found reference to a losing streak of about 21 bets so far. I am considering a low strike rate method with a good average win dividend.

Comments are appreciated.



<font size=-1>[ This Message was edited by: cameron398 on 2003-10-25 13:59 ]</font>

<font size=-1>[ This Message was edited by: cameron398 on 2004-06-15 17:54 ]</font>

<font size=-1>[ This Message was edited by: cameron398 on 2004-06-15 17:55 ]</font>

[Big Ted]

Hi Cameron
I like the idea of betting on 2 horses per race because with the right selection criteria and a good staking plan it swings the odds back my way a little.

My Numbers this week are as follows -
25.10.03 32 bets 18 collects 5.31 div 49% pot
23.10.03 15 bets 12 collects 3.37 div 33% pot
22.10.03 29 bets 18 collects 4.02 div 31% pot
18.10.03 24 bets 09 collects 4.89 div -3% pot

The longest loosing streak for the week was 6
the longest winning streak was 9 and 6 twice.

The longest loosing streak has been 10
(in 623 races)and that was on the 17.10.03 but on that day i ended up with a 30% pot after selecting a $35 winner (NR 3 no. 5)but then isnt that racing.

[Merriguy]

Hi Guys,

I like to do a bit of 'Dutching', too, Big Ted --- nice to have a couple of horses going for you in a race. Can be expensive if you don't pick the winner in your two or three :sad:

How on earth did you choose a $35 winner as one of two??? Or was it just a lucky slip! I happened to hit a $60 winner at the dogs the other night in that way. At least it makes up for the mistakes when you err the other way.

Sorry, Cameron, I know I haven't really commented on your post. Despite the best of intentions I rarely keep records --- and when I do the selection systems are so different that I doubt they would mean much.

<font size=-1>[ This Message was edited by: Merriguy on 2003-10-26 13:07 ]</font>

[cameron398]

Hey Ted. Your figures are impressive. Am I correct in assuming you are betting each way, rather than for a win?

[stebbo]

Hi Cameron,

I have an experimental system that I have recently started running on a real bank. It has been backtested over the last 20 months. It is actually a combination of two systems which will sometimes pick the same horse in the same race, but never different horses.

There have been 1800 races bet over the test period, with a total of about 2400 bets. (if both systems pick the same horse I double my bet).

The longest run of outs over that period of time was 24. However, the maximum drawdown of my bank would have been 65 units. If you're trying to figure out a bank size, I would suggest you base it on the 65 unit drawdown, rather than the 24 outs.

Hope this helps.

Cheers,
Chris.

PS. I should have stated that my overall strike rate is approx 25%.

<font size=-1>[ This Message was edited by: stebbo on 2003-10-26 19:49 ]</font>

[Big Ted]

Hi Guys
I’ll clear up a few things first
1) I only bet on 2 selections per race even though I do the ratings for every runner.
2) All races that are selected are done so by computer under a set criteria devised by myself
3) I bet the same amount on both horses no matter what price they are, I don’t have a target figure to win per race as in “Dutch betting”.
4) I bet to win only and not each way.
5) Last but by no means least I use a staking plan.

What I have tried to do is take the human element out and the above criteria will do most of that for me.
They are the 5 simple rules I follow. The first two are simple enough, the 3rd is purely greed, as if the longer priced runner gets up I want to have as much on it as possible.

The 4th is simple; I do not believe it is viable or sensible to back 2 horses for the place in the same race, just my opinion.

The 5th again is simple; a good “staking plan” is worth its weight in gold. It controls how you bet, keeps records of your bets and again takes out the human element.

Most people will tell you that level stakes is the way to go, I prefer a staking plan but it’s a personal choice.
I don’t believe that you can compare them against each other because they are two different “animals” so to speak.
What I can tell you is that neither one will make you money if you don’t pick winners.

I hope this helps

[Bhagwan]

A general rule ,is to multiply your theoretical worst run of outs by 3.5,if betting level stakes, to work out your bank size.
Remember ,
No Bank = No Chance

Get the bank size wrong & it will be purely luck only, that might save you.

Here are the theoretical worst run of outs you should expect to strike at some stage of your punting career ,depending on your strike rate average.
S/R-Outs in a row
10%-65
15%-42
20%-31
25%-24
30%-20
35%-16
40%-14
45%-12
50%-10
55%-9
60%-8

[puntz]

would it be wiser to assume not only losing streak factor, but at the end of the streak, how much $$ was lost ?

[cameron398]

Thanks Bhagwan - was waiting for a numbers guru like yourself to speak up.

Puntz, I agree with your comment. I have already put a max $$ figure on the losing streak scenario. Just trying to get some ideas as to how often it may occur.

I start getting into trouble at about 25 plus losses in a row. However, this is not a problem if you set a max loss limit to bug out at and have just had many days of good profits.

[Chrome Prince]

My advice is to mix your systems.

If you have a system which picks winners 50% of the time and one which picks them 20% of the time, you can cut your losing streaks by combining the two.

[puntz]

Cameron,
25 losses in a row, depending on the duration of that streak occuring, question goes begging to the selection method ?

[cameron398]

I agree Puntz - average win dividend is $9.00 though which seems to help greatly.

[becareful]

Cameron,

There is a "losing streak simulator" available on the net where you can put in your strike rate and get a simulation of how often you will have a losing streak of each length. Work a look:

http://mailer.today.com.au/fairfax/users/sidsid

Also you may like to have a look at a topic from some time ago which also discussed losing streaks:

http://www.propun.com.au/forums/viewtopic.php?topic=2558&forum=7

[puntz]

Cameron, so far you are still a step ahead in this string.Very good pointers in those url's.
Once you agknowledge there is going to be a streak, then it's past the stats.I think the stats. show is does happen. Good.
Once one aknowledges something, in this case, the losing streak, what's next, is it strategy or improve on selection method. And it will still exist.
Therefore, set gaols to at least re-coup your initial investment.
Gee, all or most have read it somewhere, Discipline, what else ? The first racing book I read, first words were just that, Discipline.



<font size=-1>[ This Message was edited by: puntz on 2003-10-28 15:27 ]</font>

[cameron398]

Well Puntz, here is one possible answer to the "next step". The next step is to compare the losses on the losing streaks against the profits made on good days and see how it all finishes up.

Lets say we have to survive for 8 to 10 days to make enough profits to cover a 30 plus (an approximate $1500) losing streak. Set the rule to stop at $1500 outlay lost.

What sort of strike rate would be required then to theoretically survive? Goodness this is getting complicated. But we all LOVE it I know hehe

Thanks for the link to the losing streak simulator Bhagwan. I think it confused me a little with the two columns of results though. Can you please explain this for me?

So once again to play with examples say over about 1000 bets we see a result of one losing streak greater than 30bets or $1500 so we cop the loss. On the other good days we make $8,000 clear profit.

Stats say that we should have had more than one 30+ losing streak over 1000 races.

Based upon the above tho In theory that means we could have still made a profit after suffering 3 or 4 incidences of 30+ losing streaks.

Am I on the right track here?

<font size=-1>[ This Message was edited by: cameron398 on 2003-10-28 15:29 ]</font>

[puntz]

Cameron,
I re-read this post.
There are days you say you make 8000.00 on a day. Is this Win bets or a combination of multiple types of betting?

If it's multiple, then which particular type of betting are you exempling, Win only or whatever.
If it's a Win only type of bet,s then here goes:
On a day where you make 8000.00, somewhere from there you are expecting a ,losing streak of 30, or wait till you lose 1500.00
If you made 8000.00, then keep going till you think the loosing streak will reoccur.
or, when you lose that 1500.00 limit you mention.
On 1000 races, how many days on average do you bet on, 1000 per day, or week ?
Cos 8000.00 per day * 7 days = lots !

[cameron398]

My quote from earlier "So once again to play with examples say over about 1000 bets we see a result of one losing streak greater than 30bets or $1500 so we cop the loss. On the other good days we make $8,000 clear profit."

Apologies for the ambiguity Puntz - I re-read it myself and see now where the confusion is. What I meant to express was that over 1,000 bets I observed one only losing streak of over 30 bets which cost $1,500 as a loss. The other 970 bets resulted in a clear profit of $8,000 whcih after you deduct $1,500 loss is still a nice profit of $6,500.

Having said all that - I added that you could suffer 3 to 4 lots of $1500 losses and still be in front by a reasonable figure.

I am still confused as to how to read the stats on the "losing streak calculator" for a 9% strike rate. Here are the figures

I tried to insert the figures here but could not get them to format right - so u will have to look at the link provided by Bhagwan to see what I mean.

Hey $8,000 a day would be nice hehe


<font size=-1>[ This Message was edited by: cameron398 on 2003-10-29 22:52 ]</font>

<font size=-1>[ This Message was edited by: cameron398 on 2003-10-29 22:54 ]</font>

<font size=-1>[ This Message was edited by: cameron398 on 2003-10-29 22:56 ]</font>

[puntz]

So 970 bets for a profit of 8000.
What would the total dollar amount be if it lost 30 bets ?
No need to answer, but if you stop short at 1500.00, cop it sweet, I think it should see you thru. if ya have figures of 8000.00

If I had the stuff to do this, I would look at a graph and find it's peak Gross profit per day, then be satisfied with an average.




<font size=-1>[ This Message was edited by: puntz on 2003-10-29 23:51 ]</font>

<font size=-1>[ This Message was edited by: puntz on 2003-10-29 23:53 ]</font>

[becareful]

Cameron,

In that losing streak simulator the first column is the length of losing streak (so 0 means you won on the first bet, 1=won with the 2nd bet, etc). Second column is the number if times it happened during the simulation and the last column is the number of times you EXPECT that losing streak to happen mathematically. I wouldn't worry too much about this last column as it can be a bit confusing if you are not used to probability analysis.

Plugging in a 9% strike rate and running a simulation I got these values for the large losing streaks:
26=1
27=2
30=2
31=1
33=2
29=1
46=1

Now if you decide that you are going to cut your losses at 30 then in this simulation you would have had to do this 7 times (2+1+2+1+1) as there were 7 streaks of 30 OR MORE. So if you projected profit for the 1000 bets is $8000 LESS $1500 for each losing streak you would have lost $2500 overall as the 7 losing streaks cost you $10500.

Hope this helps to clarify it.

[cameron398]

Thanks Becareful, thats the clarification I am seeking. The numbers do not look that good when framed mathematically. So a series of 1000 bets where only one 30 plus losing streak reared its ugly head would appeare to be "pure ass" !!!! Again, not want I want to hear, but necessary to digest.

[Bhagwan]

Cameron
The 31 out is related to an average strike rate of 20% ,i hope you can work it out from there.

[cameron398]

Well what do you know!! At the risk of looking like an idiot I just realised my 9% strike rate was actually one winner in every 9.4 bets on average - that seems to be actually a strike rate of 11.24% - if that is how I should be working it out???

That improves the outlook a little...

[gunny72]

I am sure these figures have been shown before but I feel it is time to air them again in the light of a lot of the recent postings. They give the theoretical run of outs that could be expected for various strike rates to be more than 99% sure of getting at least one winner (if I did the maths right).

sr 1in2 1in3 1in4 1in5 1in6
99% 8 14 19 24 29
sr 1in7 1in8
99% 35 40

Of course this analysis assumes that a punter consistently gets a given strike rate. Unfortunately, although over a period of time one might achieve a 1in3 strike rate, racing is unlike tossing a coin or a dice (which have a fixed number of faces and the chances are the same each throw) and there might be long stretches within that period where the strike rate is 1in6 or 1in2 even, but averaging 1in3 overall. Thus basing future staking on past POT has serious flaws as I am sure many have discovered in practice.
gunny