JonH comes home wrote:

---

Coup Droit wrote:

---

JonH comes home wrote:

---

JonH comes home wrote:

---

Ryan and Jay played in November, Ryan won 2 and 1 on that occasion. Got to imagine Ryan is going into this as favourite and, indeed, tennis abstract rates it 69-31 chance in Ryans favour.

---

 Now shown as 66-34 , not sure why it would change unless they added new results into the ELO algorithm. 

---

 It's the money, Jon

The change in odds shows the flow of money 

---

 Ta doesnt do it that way though , CD - its not betting odds, its the chance of winning thrown out by their elo stats - but it often ends up being similar to the odds! 

so it will only change when Elo is updated if that makes sense ?

---

 OK, sorry, my fault, never use it - I thought it was betting odds - sorry

Maybe they use betting odd sites though, in their algorithm, to reflect public consensus? 

it would be reasonable.....

This is the forecast for this tournament

http://www.tennisabstract.com/current/2023TenerifeChallenger.html

Coup Droit wrote:

---

Well, if you want ones that actually mean something, i.e. where you can hold them to account, so to speak, it's about 1.44:2.62 on most places

-- Edited by Coup Droit on Tuesday 31st of January 2023 09:45:19 AM

---

 Ha - no, |I just use it to get a sense as to how favourable our guys draws are and who might win etc...dont do the gambling thing - BTW, what does 1.44:2.62 mean in actual fact, it is just number to me!! does 2.62 translate into 1.6 to 1 in old money? That rings a bell? Which would mean the betting sites roughly reckon it is 78 to 22 percent chance of Ryan beating Jay, so more likely than TA predict it to be? 

Coup Droit wrote:

---

"Ha - no, |I just use it to get a sense as to how favourable our guys draws are and who might win etc."

That's what I mean - use a number that actually represents something, something that must - over time - be more reliable - you don't need to bet on the site

It means if you put £1 on Ryan, you get £1.44 back
£1 on Jay gives you £2.62 back

---

 I dont know actually, I think the ELO ratings (which are also surface specific remember, or at least surface weighted) are pretty reliable and I am sure Jeff ran some data to show TAs forecasts outperform the betting sites - but anyway, it is just a casual look up and it works fine for me. But thanks 

Coup Droit wrote:

---

[...] use a number that actually represents something, something that must - over time - be more reliable - you don't need to bet on the site

It means if you put £1 on Ryan, you get £1.44 back
£1 on Jay gives you £2.62 back

---

Finally!  An explanation I actually understand instead of all that boring rigmarole about algorithms & Elo ratings which is guaranteed to make my eyes glaze over & scroll past!