BSP is efficient. Is BSP really calcalated solely based on Wisdom of the Crowd?

[Tosin]

I noticed while observing and recording "last traded price" movements that sometimes when the market has been trading within a certain range for a long period, in the last second just before the market is suspended (or goes in play), a totally different price that's well outside the price range gets traded as the last traded price and hence skews the BSP far away from the volume weighted average price and price range of the pre-off market (before the final second). For example, a market could be trading around a price of 4.5 in the final minute pre-off (expecting the BSP to also be around that price) and suddenly, somehow in the last second, a price of 6.8 is traded in the last second and the runner ends up with a BSP of 6.0 even though majority of the buy/sell price was well below 6.0. What causes this to happen? Is this some form of action to keep the BSP efficient when the current price does not reflect the true price? And how is it even possible that an order that's well outside the current trading range gets accepted in the last second? From my example, if there's still available lay price within the range of 4.6 to 5.0, why would someone accept a lay price of 6.8?

[Derek27]

Welcome to the forum Tosin. The link below explains how BSP is calculated and how money on the exchange at off-time can affect the BSP.

https://promo.betfair.com/betfairsp/FAQ ... kings.html

[Euler]

BSP is actually flawed as it's only reconciled about six seconds into the race, so it actually reflects the early part of the race.

Overall any measurement in a limited frame has to be efficient in aggregate as you may have one price badly price in one direction that is counteracted by another in the opposite. So defining what is or isn't efficient is the key.

[Atho55]

Recorded the start of the last race and forwarding 1 frame at a time, BSP looks to be declared prior to the race going from Suspended to In Play so not sure how the 6s delay works.

BSP declared.jpg

4 Frames later the race was turned IP with the Total Volume value still the same.

[Tosin]

BSP is actually flawed as it's only reconciled about six seconds into the race, so it actually reflects the early part of the race.

Overall any measurement in a limited frame has to be efficient in aggregate as you may have one price badly price in one direction that is counteracted by another in the opposite. So defining what is or isn't efficient is the key.

Okay this makes sense. Thanks

[Tosin]

Recorded the start of the last race and forwarding 1 frame at a time, BSP looks to be declared prior to the race going from Suspended to In Play so not sure how the 6s delay works.


BSP declared.jpg

4 Frames later the race was turned IP with the Total Volume value still the same.

Yeah this is how I believed BSP was reconciled. Just at the point of suspension, exactly before it goes in-play.

[Euler]

That's the theory but by the time the race is suspended by Betfair it's already underway on course and people have taken positions on the break at the start of the race. Therefore it's distorted.

[Tosin]

That's the theory but by the time the race is suspended by Betfair it's already underway on course and people have taken positions on the break at the start of the race. Therefore it's distorted.

Oh I see. I understand now

[Derek27]

A simple solution would be to put the race in-play as the last horse goes into the stalls. It's not really necessary to suspend bang on time.

[Atho55]

Is this to capitalise on a bet placed before the off at "Take SP"

[rik]

A simple solution would be to put the race in-play as the last horse goes into the stalls. It's not really necessary to suspend bang on time.

what if there are issues with the start, have to suspend again, will mess up some automations badly
uk racing very rarely is suspended late enough for people to take much of an advantage, some other countries like south africa, america are a lot later and maybe not such priority as turnover much less
australian racing used to be a big problem until august but they improved their feed a lot

[Atho55]

This is from a race last night. Pic 1 is the actual race off time with both the BA timer and my PC time visible. They are about 3/30ths of a second out of sync.

Picture 2 is a snapshot taken of the first frame that BSP is available with the market still suspended but the clocks have moved on a few seconds.

There does appear to be monetary changes between Pic 1 and Pic 2.

So is this BF actually knowing the exact start time and using those few seconds to reconcile SP or others taking advantage of that window and placing money in the market?

Race Off Time.jpg

BSP declared.jpg

[rik]

did you have bets in the market yourself that were cancelled or matched at SP at the start of the race?
For me the suspension is longer the more bets i have in the market that have to be cancelled or matched at the start of the race

[Atho55]

No bets in the market. Was curious when Euler responded by saying money entered the market between Actual Off and BSP being declared. He is correct at least in this instance. If the gap is circa 3s then BF reconciling all the SP bets to declare SP in that window seems reasonable. 3s is also long enough for someone to decide NOT to put the market IP for whatever reason. I don`t think you can assume that every instance is the IP controller not doing his job.