Forum

General discussion about Bf Bot Manager v3 software for Betfair, Betdaq and Matchbook betting exchanges
By dn6789
#25059
UK Horse Racing with SQ seems to work - 2 days in a row >>> so a check market there

but UK GH - no cigar as yet - going to start afresh as 2 losing days
Attachments
uk.PNG
uk.PNG (69.03 KiB) Viewed 920 times
By dn6789
#25060
still have not sunk my teeth into more form orientated UK things as yet - but I am just playing tweaking - within what I have.

I watched a few UK GH races -- the markets are different than here so I actually think if I go to the form side - there will not be a lot of bets because I will filter the crap out of it
By dn6789
#25061
the one thing about SQ betting that I like .... it is not about any one race - form means nothing - the market dictates where things fall -- we just find our little niche within that and take our profits

yesterday I was watching it and it bet on GH 2 at like 6.50 - personally I never would have bet on the dog - but it wins by 6 lengths and I thought to myself - that is why I build SQ --- to completely discount form study LOL
By dn6789
#25071
so here is where I am at with UK Horse Racing

I took your original one made some changes and over 2 days it is 2/4 +$26.50 ... ok that is fine right

then I did a new modified one how I would build it and over 2 days it is 8/14 SR 57% +$85.65

for someone who knows nothing much about UK racing 10/18 53% +$112 .... I will take it

not complicated but again it just shows you that you can win at this with very little effort
Attachments
uk sq.PNG
uk sq.PNG (40.25 KiB) Viewed 912 times
By dn6789
#25076
$88.85 9/16 for mine

$29.70 3/6 for more original

we know from looking at GH these are running slightly higher than the 40% I see over there -- but the races perhaps are even more filtered here than in GH which is giving us the higher strike rate

if this continues to see high SR will probably move to a more advanced recovery/staking option - but right now just using low stakes

again the thing that appeals to me is it is based on SQ technique - so form is irrelevant - it is just figuring out which races run according to the expected results more often and avoiding those that do not ..... this is where our edge comes from .... other punters are assessing the race given the hundreds of races a day where things are all over the place -- where we are much more selective which probably is a good thing.
By echelon
#25086
Hi DN,

That's a great improvement on the original :shock:

It appears from what you're saying that it's including the SQ which makes all the difference.

I've looked back at the favourites in my bets from the original and have built a SQ based on the patterns I've seen as:

1,3,4,2,1,5,1,2

So that's now in Sim and we'll see how it fares against the original...
By dn6789
#25087
last night original down to 14.70 3/9 my modified one 12/24 100.50

as far as SQ pattern goes it really depends on how well you filter the race

if you specialize in 7 horse fields and reduce the races down 20-50% then really your going to see alot of 1 2 3 or 4

and only rarely a 6 or 7

I am still learning UK Horse so I will know more when I see few more days betting - but so far the principles of AUST GH SQ seem to be working
By echelon
#25094
An interesting start for my SQ test as it doesn't seem to be quite as good as the original.

At least both are showing a nice profit to flat £1 stakes :)

It's early days so I need to compare them both for a few more days I think.
Attachments
Horses_Test_4.JPG
Horses_Test_4.JPG (38.84 KiB) Viewed 883 times
By dn6789
#25099
more data - more tweaking - more filtering then figuring out the staking

this is what you and I both need to do with this UK stuff -- it is not easy to win betting - we have to develop the best strategy we can and it is too early to say it is solved.

AUST - I feel very comfortable in saying this is solved.
By dn6789
#25100
and when I say solved - what I mean is the strike rate is high enough with staking to win everyday

solved does not mean winning 100/100 bets - it is impossible to do that.
By echelon
#25104
Now this is really annoying...

Great results in Sim (and SQ1 performing better than the original today - no changes have been made).

However, losses in live!

If the odds in Sim are delayed by up to approx. 2 minutes, is it worth entering bets earlier in the live strategy in an attempt to align with Sim?
Attachments
Horses_Test_4_Mon.JPG
Horses_Test_4_Mon.JPG (39.15 KiB) Viewed 854 times
Horses_Test_4_Live_Mon.png
Horses_Test_4_Live_Mon.png (13.23 KiB) Viewed 854 times
Compare_Test_4.png
Compare_Test_4.png (39.45 KiB) Viewed 854 times
By woweee
#25106
Hello echelon

I have been using your Horses Test 4 system. I have made the following changes.

1. Time to Bet 0 seconds.
2. Fav must be priced at 3 or more. Does mean you can win and still make a small loss.
3. Highest Matched amount. Don't think this makes any difference.

I have also included a few dog strategies. Which have managed a profit over the last 4 days.

I found that the Australian dogs have very little money matched. The money only arrives seconds before the race start.
So have the time to bet set to 0 seconds. The Dog strategies seem to work better on the Australian dogs.
Trap 1 has the most winners. Followed by Traps 2,3 and 4. Especially on the Australian races. If the Fav is in trap 1 it stands a very good chance of winning.
Strategy 2 had a bad day today. It seems to work better on the Australian races.
All strategies are staking 0.10p
No recovery.
Dog strategies are betting on Aus and GB.
Horses just UK. May try US tonight.


I have attached the strategy file so members can have a play.
If you can see how to improve it. Please let me know.

I have tried adding SQ. Just can't see how it can work. I don't think many member's do.
It was a total failure. So unless more info becomes available I will leave it well alone.
Attachments
Test.png
Test.png (170.89 KiB) Viewed 852 times
(5.05 KiB) Downloaded 56 times
By echelon
#25107
Hi Woweee,

Great progress! Many thanks for posting your strats so far for others to work on further improvements.

I'll run your horses test 4 against the 2 versions I'm currently running and compare the differences.

Setting bet time to 0 seconds, do you find that you're missing some dog races? The issue being is that if you're running a lot of strategies and have a lot of markets - e.g. I have dogs, horses and some football markets loaded, then it takes time for each strategy to scroll through the markets and that delay can prevent you from entering any bets until after the race has started and in the case of a dogs race the market will be suspended when the race is on.

I've added a screenshot to show how I'd start to look at SQ patterns. I've taken the bets from Horses Test 4 and filtered them to show the winners only.

Now we can look under the favourites column to see if there are any obvious repeating patterns - I've circled some of the obvious repeating ones.

Next I would look at the proportion of each favourite winning:

Fave 1 won 11 times
Fave 2 won 8 times
Fave 3 won 3 times
Fave 4 won 3 times
Fave 5 won 2 times

Hence, we know that in any sequence we try, there should be a lot of 1's and 2's and a handful of 3,4 and 5.

We could create a sequence of just 1's and 2's, though we'd miss out on some of the higher odds winners when they're picked up as part of the SQ.

Now we need to combine what we know from above and create several sequences, so that we can test them all in Sim to see which are profitable.

Here's an example which has some of the ringed patterns, plus others such as 1,3 (which appears 3 times)

1,1,2,1,3,2,2,5,1,4,2

Just create lots of these and test them out with the Horses Test 4 selection criteria. Always start them with the favourite most likely to win (i.e. 1).

The idea is that patterns which we see repeating are likely to happen again. The reason for this is that our bets are already pre-filtered to very specific conditions. Just betting on Fave 1 in every race would show too much randomness in the winner sequence.

The sequence is set to restart on a win and to restart at the end. So, we only progress further through the sequence when our bets are losing. As Fave 1 and Fave 2 make up 70% of the winners, then it's a good idea to have these at the start of the sequence. The rest of the sequence can be scattered around and include a few of Faves 3, 4 and 5.

If we hit a run where Faves 1 and 2 are losing, then we progress through the sequence, eventually hitting winners from those 3, 4 and 5 Faves at nice high odds :D
Attachments
SQ_Horses_Pattern.png
SQ_Horses_Pattern.png (126.87 KiB) Viewed 845 times
Last edited by echelon on Mon Aug 23, 2021 8:33 pm, edited 1 time in total.
By woweee
#25109
Hi echelon

Wow a lot to take in.

Don't think I'm missing any bets.
These are my monitor settings.
Attachments
Monitor.PNG
Monitor.PNG (19.48 KiB) Viewed 844 times
By echelon
#25110
Ok - you have a nicely optimised set up and without overloading on markets and strategies, I can see that you should be fine.

Sometimes, I have thousands of markets active particularly for football strategies (You can have match odds, BTTS, all the overs/unders goal markets etc.), and I've experienced bets being missed.

Just to clarify for anyone wondering how to set up sequences - it's the selection condition as shown in the attachment.

An short version of sequence is SQ which you'll see mentioned in other threads
Attachments
SQ.JPG
SQ.JPG (52.27 KiB) Viewed 843 times
By woweee
#25112
Just thinking out loud.

Instead of a sequence that places only 1 bet on a race.
What if you doubled up.

What if your first step bet on Fav 1 and Fav 2. and after a loss you bet on Fav 2 and Fav 3.
And still restart on a win. And return to the start when you reach the end.


This data is from your earlier post.
1 W Bet on Fav 1 and 2
1 W Bet on Fav 1 and 2
1 W Bet on Fav 1 and 2
2 W Bet on Fav 1 and 2
1 W Bet on Fav 1 and 2
3 L Bet on Fav 1 and 2
2 W Bet on Fav 2 and 3
2 W Bet on Fav 1 and 2
2 W Bet on Fav 1 and 2
5 L Bet on Fav 1 and 2
1 L Bet on Fav 2 and 3
2 W Bet on Fav 1 and 2
1 W Bet on Fav 1 and 2
3 L Bet on Fav 1 and 2
5 L Bet on Fav 3 and 4
1 W Bet on Fav 1 and 2
4 L Bet on Fav 1 and 2
1 L Bet on Fav 3 and 4
1 W Bet on Fav 1 and 2
2 W Bet on Fav 1 and 2
4 L Bet on Fav 1 and 2
1 L Bet on Fav 2 and 3
1 W Bet on Fav 1 and 2
3 L Bet on Fav 1 and 2
4 L Bet on Fav 2 and 3
2 W Bet on Fav 1 and 2
2 W Bet on Fav 1 and 2

Thats 16 wins.
11 losers.



I think you would have to use two strategies.
One to bet on fav 1 and 2.
And another to bet on Fav 2 and 3/
Last edited by woweee on Mon Aug 23, 2021 10:22 pm, edited 1 time in total.
By dn6789
#25113
my UK GH lost $50 and the original horse one lost $10 and the modified one was about flat.

for me this is not really acceptable - as I expect every day to make money with every strategy -- even if it is $10 or $200 they are all designed and updated to be the best they can

this tells me that the variables for UK at their core are different and I feel like my filtering needs updating.

one other problem is I am not awake to see these live most of the time so that frustrates the crap out of me.

something else I mentioned yesterday that I do not think people really understood.


say I specialize in 1400m horse races 6-7 horse fields Fav 1 drawn barrier 2,3,4 Fav 2 drawn 6 or 7

Fav 1 price $2.25 Fav 2 price $3.80

does this mean that Fav 1 wins the race 100/100 times ?

No that is why even a small SQ to randomize the result to a few runners can help you win rather than just bet Fav 1 every single time

or you could say OK I want to be a Fav 4 bettor - what conditions does Fav 4 need to win the race

you gotta figure out this stuff -- it does not work if you just allow bets on every race where the results do not favor the bets.

(you are either going to understand what I am saying or not - and that will be the difference between success and failure)
By echelon
#25116
dn6789 wrote: Mon Aug 23, 2021 10:21 pm one other problem is I am not awake to see these live most of the time so that frustrates the crap out of me.

something else I mentioned yesterday that I do not think people really understood.
That's exactly why I like to concentrate on UK races being based in the UK, though as you've found out, there appears to be a different driving force behind the results between AU and UK - could it be connected with a different style of handicapping?

I think I understood exactly what you meant, and I've noticed that you've made the same point in slightly different ways before.

I believe we need to do the following:

1) Download the Results data from BFBM
2) Identify a combination of some or all of these: Distance, Grade, Favourites, Traps (Boxes), Prices of runners
3) Each line in the Results data gives us a winner, so we need to filter our data as in 2) and look to see if there are any recurring themes.
4) Modify the filtering criteria to also capitalise on those recurring themes.
5) Create a strategy in Sim to include filters from 2) and 4).
6) Monitor SR and DD and adjust if required.
By dn6789
#25117
yes - you need data to tell you what to make

then you need to activate it to bet - and see if your results match the statistics -- then filter out races - runners and lower your bet numbers to the most high quality you can - which will give you better strike rate and make you more consistent

next look at your staking and see what matches to the results your seeing.

this is not going to happen in one day - you need some experience which is why it is no good just reading this - you gotta try it and see what happens -- then adjust.

anyway I am out till later now - got errands to run
By dn6789
#25125
I have been thinking about this given that the UK seems more complex especially GH

the reason to me is the price .. in AUST races range from 1.18 to 300/1 where in the UK alot of times the field is 3-9 4-10 a much more compressed market is what I am saying

not sure if you guys are familiar with pokertracker software - but I have a feeling we need a HUD like that on our screens as we need a lot more information -- + we are going to need form filters

for me yes this is still part of what I want to do -- but because AUST/NZ delivers guaranteed profits everyday pretty much - my mornings are spent keeping my strategies smartened up there - than pouring through data here.

I really believe that this sort of betting is due for an update - based on a data approach but the tools to do it simply are not at a level that we really need.

Again as I searched online for this sort of thing - you can just see how out of touch the so called experts are in terms of day to day racing and making a profit.
By dn6789
#25126
plus the 5-6 people really taking an interest and adding something back have all said to me - that the tracks etc are different in so many ways as well.

it is on the agenda but not something I am going to kill myself to achieve this week for example .... I just want to prove that I can do it but more at my leisure sort of thing.
By dn6789
#25127
if I can pass one thing onto to you about betting races ...

I found that building a strategy per runner is better -makes more money - and lower DD than by trying to say

OK - I am going to bet 2 runners per race vs fav 1 as an example

if you say to me why is that - I think it has to do with the price .... to make it profitable you really need higher prices than what you end up achieving - because you have a double bet per race - and it is a slow bleed despite winning days here and there ... I kind of already tried it for some reason -- and filtering races and runners works so much better

good luck
By echelon
#25135
Hi DN,

You're right about that price compression.

In AU races 7th and 8th favourites typically win at least once in a day's racing and sometimes with massive odds.

Whereas the UK favourite with the largest odds is 6 and I've sometimes seen that winning with odds as low as 7.0 !

I believe that there are patterns in the UK GH data which can be exploited. It may be that we need to be betting on the place market to boost SR at the expense of profit or we could be laying instead of backing.

Given that UK odds are generally low, then creating laying strategies should work well,

I've done some analysis of UK GH Win market races this year and there are ways of turning a reasonable profit (even by backing!) looking at non-form data.

The issue with using form data is you mentioned is that it's not available in BFBM. If it was, then average speed and previous finishing position when running the same distance would be useful to know!

I'm struggling to see why building a strategy per runner is more profitable.

If I have Strat Combined which looks a Fave 1 in traps 1 and 2, why should this be different than separated Strats F1/T1 and F1/T2 ?

If we're using a level stake then the results should be pretty much identical - apart from perhaps the odd price difference caused by the delay between betting from the separate Strats.

Aha - However, if we're using any sort of fancy staking then built up clumps of losses in the combined strat could increase our stakes much more than in the separate strats. (thinking out loud here...)