Current Match Engine

mike778
Posts: 436
Joined: Sat Apr 23, 2016 10:55 pm

Re: Current Match Engine

Postby mike778 » Sun Sep 23, 2018 8:38 am

I'm not sure how viable it would be in terms of the code and audit logs and so on. But would it be useful if the Devs had a look at a brilliant ball that was rated 10 and was hit for a boundary and a crap ball that took a wicket and see why these events happened. If the brilliant ball was slightly less good (8) then would it still have been a boundary. I suspect this isn't possible but depends on how the ME works and results are logged ...

Is it possible for the devs to run a simulation (1000 balls) with all variables the same apart from the bowlers and batsmen. Take two bowlers who are identical types (same traits, stamina, skill distribution) but give one of them slightly higher skills and see what comes out.
Leader and supreme commander of Chester's finest (Chester Ducks) cricket team -
Season 17: cup runners up.
Season 18 : 2020 runners up
Season 18: SOD runners up (bit of a pattern emerging here sadly)

rtj45
Posts: 3276
Joined: Sat Apr 23, 2016 12:31 am
Location: Dorset

Re: Current Match Engine

Postby rtj45 » Sun Sep 23, 2018 9:09 am

bumpuss wrote:
MOD-tonk6464 wrote:I've watched a huge amount of cricket in real life and I've observed the following: Poor balls do take wickets at times. However, good balls take wickets more frequently (particularly jaffas). The current Stumped ME, whilst good in some aspects, doesn't reflect this and I think it would be better if it did.


Thats fine, what is poor ball rating - 1-7? 1-9?

Good ball rating just 10/10?

Like I said at the moment very few balls are in the 8-10/10 range...bump up the that and you will have more wickets in that bracket (without actually changing anything in the ME - just a visual change).


The figures are based on percentage of balls taking a wicket, its not the sample size thats the problem. Nice try though.
Manager of West Row Warriors
Winner of all available titles in Season 13 (Dev, T20, SOD, Cup, International)
Breeder of Peach #1 bowler in the world.
Former world #1

bumpuss
Posts: 205
Joined: Sat Apr 23, 2016 12:30 am

Re: Current Match Engine

Postby bumpuss » Sun Sep 23, 2018 12:05 pm

rtj45 wrote:
bumpuss wrote:
MOD-tonk6464 wrote:I've watched a huge amount of cricket in real life and I've observed the following: Poor balls do take wickets at times. However, good balls take wickets more frequently (particularly jaffas). The current Stumped ME, whilst good in some aspects, doesn't reflect this and I think it would be better if it did.


Thats fine, what is poor ball rating - 1-7? 1-9?

Good ball rating just 10/10?

Like I said at the moment very few balls are in the 8-10/10 range...bump up the that and you will have more wickets in that bracket (without actually changing anything in the ME - just a visual change).


The figures are based on percentage of balls taking a wicket, its not the sample size thats the problem. Nice try though.


What figures?
Are your better bowlers taking more wickets compared to your worse bowlers? The better bowlers will have better bowling ratings.

Ravenprince
Posts: 432
Joined: Sat Apr 23, 2016 12:31 am

Re: Current Match Engine

Postby Ravenprince » Sun Sep 23, 2018 2:30 pm

@bumpuss
- The ball rating data shown in the commentary is a derivative of what the ME uses to calculate outcomes. A higher rated ball is what the ME determines to be a better ball, which is harder for the batsman to play. The data is already bumped up the way you suggest, most common ball ratings are 7 and 8, 10s are rare but so are 0s and 1s.
- Accepted crap takes wickets, but over hundreds of games over a couple of seasons, lower rated balls took wickets at a similar average to the better ones. Pretty much everything above 2 gets wickets at similar strike rates, some lower quality balls get wickets at a faster rate some than higher quality ones (I think 4s or 5s get wickets almost 10-20% more frequently than 9s - these occur at similar rates in a game). James and I collected this data and it is posted on the forums.

The summary is better bowling does not really translate into better outcomes. If you are really interested, there are a bunch of posts which you can go through to catch up on the discussion.


Return to “Match Engine”

Who is online

Users browsing this forum: No registered users and 2 guests