Why Systems don't work

[thekey]

A Quote from Erik Hoffer

"In times of change....
learner's inherit the earth,
while the learned
find themselves beautifully equipped to deal with a world that no longer exists."

While this was aimed more at stockmarket type investing it applies also to horse racing.

Basically, I think, it is more important to know how to apply the past (performance history) to the future than it is to concoct ways which would have made money in the past, if only we had known about them beforehand.

[becareful]

Ahh, but "History always repeats itself" - don't have a clue who said it.

In the stock market there are a lot of people who make money based on following rules developed around trends, etc (technical analysis as opposed to fundamental analysis). These rules are based on historical analysis of trends and do work (well some of them work anyway - I personally know one person who was making a good income this way until the stress started getting to them).

There is no doubt it is possible to develop a system based on past results BUT there are a few things to remember:

1. You have to do it yourself - buying a system will never work because once a few people start using the same system it will stop working.
2. The rules you come up with must work consistantly month after month (no good coming up with something that shows a huge profit 1 month but losses the rest of the time). If you can find rules that show profits for at least 90% of the time in the past (with several years data) then there is a very good chance it will continue to work in the future.
3. The rules should be as simple as possible - eg. bet on dead or better tracks only is OK - bet on heavy if it is first Monday of month is NOT (this is a specific rule to fit past data)
4. If you come up with a system that works DO NOT SELL IT - you will make more money using it yourself!

Sounds like you already have one.

[becareful]

Still refining it but it has made a profit for 11 out of the last 14 weeks. :smile:

[thekey]

I agree becareful,
The point is a rigid set of rules won't work, you must be flexible and constantly able to adjust for the circumstances.
This does not mean historical data is useless, it is in fact very helpful and a good pointer to the future. But even if it were possible to run a race of exactly the same field under the exact same conditions it doesn't mean the same horse would win.
When you get your head around that you will be on the right path. I Hope!

Another thought on systems- the marginal selections, when back checking you kno the results and therefore will probably include the winners and leave out the losers, will this happen when you have to put your dollars down?

I will have more to say about this in the future.

Any chance of getting some tips for saturday?

Food for thought: When looking at past data, ask yourself "Which of these outcomes are more likely to re-occur in the future?"

Some things repeat, others fade away. Work out which things are going to repeat and which are going to fade away and you have found something worthwhile.

[thekey]

Yes Freddy I'll try to post some selections.
I'll even provide a suggested value price!
I guess this will test out my theories.

[becareful]

Another thought on systems- the marginal selections, when back checking you kno the results and therefore will probably include the winners and leave out the losers, will this happen when you have to put your dollars down?


I don't have this problem as I do all my back checking on the computer - I adjust my criteria and it applies that criteria and gives me the result - therefore I don't manually include or exclude marginal calls. Of course when I am placing my bets I do have the option to override the system but I am learning not to! Have had several instances where I missed out on big collects because the system told me to bet but I thought the horse had no chance only to see it come in.

With regards to adjusting the system I am constantly reviewing it but am finding that the changes I am making are really now just very minor fine tuning.

My system is not trying to pick the winner in every race - simply identify horses that are good value (eg. $15 when they should be $10). I am finding that my "rigid rules" are actually surprisingly good at identifying those horses (better than I expected actually). Therefore it doesn't matter that no two races are ever the same - as long as there are value horses in the field that are mispriced on the tote I will be on the right road!

[thekey]

My Top-Rated Runners for Sat 29/6

Ipswich
1/2 Gullcatcher $5.00 6
2/6 Nova 3.20 10
5/3 West Ward 5.20 6 (should get soft lead)
6/3 King Lotto 8.20 4 (very open 8th rated is 14.90)
7/8 Flying Sparks 6.40 5
8/1 Darcy D'or 2.00 10

Rosehill
3/1 Timidity 3.10 10
4/6 Charmeuse 5.90 5
6/11 Zabarra 4.60 7
7/3 Magic Feather 3.30 10
8/4 Oakfield Duke 2.80 10

Will only count those that are available at value prices in results. Staking will be to rated prices(between 3 & 10 units)

Have ignored Flemington for obvious reasons!

[thekey]

That was not pretty.

Result 6 runners were available @ the right? price.

Outlay 45
Return 0

Just one of those days.

Erik Hoffer must be to blame.

[Equine Investor]

There are two things to consider with any selection method...

(1) The best horse in the field does not necessarily always win the race, in fact I would suggest that the ratio is a lot lower than what one would think. So if your selection method picks the best horse on class, form, strike rates, barrier position etc ; you may not necessarily get the winner even though your research says that it is.

The varying reasons for this are luck in running, suitable pace, barrier draw, the jockey riding the horse exactly to the very best to suit the horses method of running and being able to give it the run of the race.
The list goes on.

(2)No rule should be set in concrete.
For example you may include any horse which has a 50% place strike rate and discard any below this. However, there may be another horse in the field which has been running below it's best because of bad luck, bad barrier positions etc and comes out and wins a metro race by 5 lengths. Now that horse may only have a 20% place strike rate and so it doesn't qualify - and yet it should well be included.

So my summary is this...

You must have rules, and guidelines. But each race is different, and each horse is different, therefore you must be objective and bear in mind your rules, but look at what in the field may be an exception to this rule.

Just my thoughts.

A lot of those horses weren't the best in the race. I think that's where the problem started. Bad luck is not to blame so either blame Erik or yourself.

<font size=-1>[ This Message was edited by: freddy on 2002-06-30 10:51 ]</font>

A few observations on these horses:


Ipswich
1/2 Gullcatcher - racing against rubbish so oods weren't good.
2/6 Nova 3.20 - 7yo country horse?
5/3 West Ward - hasn't been finishing his races off.
6/3 King Lotto - Doesn't win often against quality opposition
7/8 Flying Sparks - ran a shocker LS.
8/1 Darcy D'or - rose 5.5kgs in weight

Rosehill
3/1 Timidity - Honey Ryder was down in class big time.
4/6 Charmeuse - Up in class, down 400m in distance.
6/11 Zabarra - won like a good thing
7/3 Magic Feather - 2nd up risk after 14 days, poor dividend. Has won 2nd up before but beat nothing
8/4 Oakfield Duke - due for a defeat, poor odds, wide barrier

8/10 had several things against them.

[hermes]

I didn't get to post my selections this Saturday but I did poorly like thekey and also testarossa, placegetters and others. My "sure things" were:

Rosehill

Race 3 #1 Timidity
Race 4 #4 Mr Attorney
#7 Kerouac
Race 6 #Etoile Bleue
Race 8 #Oakfield Duke

Vic Park

Race 4 #1 Zedimbi
Race 6 #1 Ex-Files

I ruined Mr Attorney - otherwise a good win - by backing Kerouac in the same race - couldn't separate 'em on my selection method.

Crash! A bad day. I'm prepared to put it down to a batch of bad selections rather than just bad luck and then try to learn from it. In hindsight now I can see the shortcomings of each of them. Glib selections, I think. Or lazy selections, I tell myself. Oakfield Duke, for instance, due for a loss, poor barrier, etc. Should have looked a lot closer.

Timidity was the first horse in recorded history to meet ALL my selection parameters - miserable.

I really thought I'd picked the best of the best horses this Saturday. Most of 'em lost. Most of 'em unplaced! In most cases it was because they were not, in fact, the best horses. I was wrong. So
a bad Saturday all round, but there were good lessons to be learnt. The best horse doesn't always win the race but if I lose it is still much more likely that the best horse did win but I didn't pick the best horse.

Conclusion: Plenty of room for improvement in my selection system!

Advice gratefully received!

Hermes

[thekey]

Don't we all have 20/20 in hindsight?

Gullcatcher: didn't even think about at the odds

Nova : carried 10kgs less than last start and had no luck in the straight I thought it was a good bet @ 13/2. My main concern was with the distance. It has won up to 1800 but that was a while ago.

West Ward: SHE was never considered at the price

King Lotto: The price thing again

Flying Sparks: chose to ignore last start for the base rating because he drew a nightmare barrier and was snagged straight back to last. (note Awesome Weather was rated 0.5kgs below and was good value, but not included because I said I would only post top rated)

Darcy D'or: yes up in weight 5.5kg but down in class about the same (OHP Rand to SCON Ipswich) has carried big weights well before, probably didn't quite get this distance with this weight. I would say the 2.00 I quoted was too short perhaps 3.00 which would have meant a NO BET.

Timidity: think you'll find it was down in class just as much as Honey Ryder and was fit and in form compared to other horses with a query on their fitness. Although the winner did have a couple of barrier trials.

Charmeuse: too short

Zabarra : very frustrating when the best price is half a point below what you want but you have to draw a line somewhere.

Magic Feather : too short thankfully as it bled and is now banned

Oakfield Duke: due for a loss ???? perhaps didn't penalise heavily enough for combo of weight rise and bad barrier. Perhaps the barrier beat him though as he crossed them pretty comfortably but maybe used to much petrol.

Freddy, let me know when there is a horse with no negatives running so I can get on at 1/2.

Honey Ryder - dropped from a Listed Race to a restricted race for F&M with 2 metro wins
Big drop.

Timidity - Racing in similar company.

With good horses like Zabarra $4 is fine. Don't know why you wouldn't take that.
May have been racing against older horses but most of them aren't world beaters. Also Zabarra had the highest $$$ per race, dropping 3.5kgs and ran good time last start.

I suppose you would've knocked back the $2.80 about Boreale - Nothing wrong with $2.80 if the horse is good and the opposition is ordinary.

[thekey]

Freddy

How about we see some rated prices from you!

Just for interests sake my ratings had 7 winners from 11 races in the top 3 so I can't be too far off the mark. 2 of the others were 4th and 5th respectively, one paid more than $21 (rated $25) and the other returned to its best after being out of the winners circle for 11 months. Two winners, rated 2nd were available at what I would consider good value and would have been enough to show a profit if backed.

I constantly review my selection and pricing technique and where I apply bonuses/penalties (note horses are considered as individuals when doing this), this is the only way to learn.

Last Saturday it was 11/14 winners in top 3 with value winners (top rated only) Heptonstall, Badgers Wood, Satashi and A Deena.

Using the same staking method the outlay was 59 units for 225 return. Profit 166 units. Puts the 45 lost yesterday into proportion?

15/6/02
5/8 in top 3
only 2 top rated available @ price
20 out 61 in Profit 41

I don't really care whether you believe this or not, but perhaps you could check some old posts of mine where I provide complete ratings for the Toowoomba Cup and Weetwood. Neither winner was top-rated but both were over the odds.

I believe yesterday was just a bad day, besides it was ONLY 6 selections.

I shall finish my contribution to this discussion with another quote, not sure who to attribute this too.

"If people learn from their mistakes, does this mean that people who have never made a mistake, don't know anything?"

Its easy picking winners if you have 3 selections per race.

[becareful]

On 2002-06-29 21:38, Equine Investor wrote:
There are two things to consider with any selection method...



EI,

I agree totally with your first point (best horse does not always win the race) - in fact as I generally only back $8-$20 runners I would probably say my system depends on this! Certainly the best horse on paper often doesn't win for the reasons you mention and a few others. One of the big problems for favourite backers is that many horses probably don't win because of things not known to the punting public (eg. training strategies where the horse is being prepared for a particular race and trainer gives instructions to rider not to ride it too hard, also things like minor injuries in training, etc). Also, like people, horses just seem to have bad days for no real reason! This is why I tend to concentrate on looking for value horses that can win if the better horse(s) fail for any reason.

Also generally agree with your second point (no rule set in concrete) although I am forming the opinion that I should not exclude horses that get included by my system but I can include ones that miss out if there is a good reason for it.

[thekey]

If its so easy quit your job and go pro.

[The Catparrot]

On 2002-06-30 15:38, freddy wrote:
Its easy picking winners if you have 3 selections per race.


I don't agree with that Freddy.
Ratings put horses in order of their winning chances. It does not mean that even one of the horses is any value and should be backed.
Good ratings long term will only average about 60% from the top three if every race is rated.
Just for an experiment Freddy, and you don't have to post it, pick your top three for races of your choice, decide which if any should be backed and how they should be staked and see how you go.
The Catparrot.

I don't need to back 3 horses. One is enough.
If you back 3 horses per race, in the 25% of races where none win you lose triple the amount! pass on that.

If you back 3 per race, and the avg winner is only $4 that's only a break-even situation based on a strike rate of 75%.

The only way for this to be profitable would be if your avg dividend was higher, whilst still maintaining the strike rate at 75%.

I'd rather pick 1 and have a strike rate of less than 50%.

[The Catparrot]

Accept everything you said in that last post Freddy. But I use ratings. My top selection might be assessed by me at $4, $3 available, no bet. My second selection might be assessed by me at $6, $4.50 available no bet, whereas my third might be assessed by me at $6.50, $12 available. So it is a bet.
The Catparrot.

[Placegetter]

On 2002-06-30 12:47, hermes wrote:
I didn't get to post my selections this Saturday but I did poorly like thekey and also testarossa, placegetters and others.


Thanks for the mention hermes :?

I was fairly clear on the fact we were betting for value last week and Smiling Eyes in particular was the longest priced qualifying runner I could find in Melbourne.

If anyone ever puts money on something I selected.......You idiot. Do your own research. I have already written last week off as a crap day. I think I'll paper trade for a week. Just like horses, we go through peaks and troughs, it's important to realise this and take a week off if necessary. Well I think so anyhow.

Placegetter

[quote]
On 2002-07-02 14:50, Placegetter wrote:
[quote]

"If anyone ever puts money on something I selected.......You idiot."

That's a vote of confidence.

[Placegetter]

Whats your point?

Who needs a vote of confidence anyway? Your bank balance will tell you whether you should be confident or not. There are only three tipsters on this site I would even read selections from, let alone note them in my black book.

I stand by my statement. If you follow my tips, or anyone else's on this site, get out before you are eaten alive. I only bet on one of my tips from last week (Jilina Blue, fourth at Ipswich), so how could you possibly know what final factors I look at before putting real money on it?

One regular contributor to this site bets ONE MINUTE before the jump, so his selections could only ever be horses to consider.

Then there's price or the dirty "v" word to take into account. Just like trainers don't enter every race to win, I don't bet on every race to win. As long as I play the percentages right, in the long term I'll be out in front.

What do you think this is? A place to log onto every Saturday morning, write down a few tips then head off to the track to secure the best price and collect your fortune. I know that place, it's called Neverland.

Placegetter

PS. Correction, I also bet on Carlton at 13/1 to beat Brisbane by 15.5 points or more.

_________________
Pick it to win, back it to place.

<font size=-1>[ This Message was edited by: quapi on 2002-07-02 22:20 ]</font>

[quote]
On 2002-07-02 15:32, Placegetter wrote:


"Who needs a vote of confidence anyway?

I stand by my statement. If you follow my tips, or anyone else's on this site, get out before you are eaten alive."

Tell us what you really think.

[Placegetter]

I don't think it would be necessary to do that, otherwise I might be removed from the forum, and heaven knows how awful that would be.

You guys not having me around to rely on for tips and great advice! Just how would you survive?

Here's a tip. Your next bet in Melbourne should be in about 12 weeks.

Placegetter

_________________
Pick it to win, back it to place.

<font size=-1>[ This Message was edited by: quapi on 2002-07-02 22:21 ]</font>