Future of the Formula BOINC

Bug reports and suggestions
User avatar
sebastien
Steering committee
Posts: 112
Joined: Thu Jan 01, 2009 2:59 pm

Future of the Formula BOINC

Post by sebastien »

I am very grateful to UBT -Timbo and Fardringle for their help in running Formula BOINC.

There are 73 teams registered with the FB but only about ten actually participate in the FB. I am primarily responsible for this state. I don't have enough motivation and time to properly manage the FB.

So, I see two solutions:
- Closure of the FB at the end of the year.
- Modification of the FB. (For example, the change could be that teams must register to participate in FB)

The first solution is technically simple, but it will be really heartbreaking.
The second solution is more difficult. I will have to delegate more management to other people. These people should come up with new unifying ideas to make FB more attractive.

So if there are people who are motivated to run the FB next year, you can discuss it here. I will just give my opinion on the technical feasibility. I will also code any modifications that might be useful to these people for the management of the FB.

If no one is motivated, the FB will be closed at the end of the year.

I saw motivated people including UBT - Timbo and Fardringle. So I hope there will be a FB 2022.
UBT - Timbo
Site Admin
Posts: 243
Joined: Wed Nov 23, 2016 1:21 am
Contact:

Re: Future of the Formula BOINC

Post by UBT - Timbo »

Hi Sebastien

Thanks for your open and honest opinion about the state of FB.

It seems true that many teams are not interested in taking part - this is a shame as some projects will not benefit from the "number crunching ability" that is possible from having this all-year long Challenge.

Perhaps some extra "marketing" is required to get more members and their teams involved (or to pitch to existing teams./members to "fire them up" and have them return to FB?)...Maybe there needs to be some better "rewards" provided to teams and their members - maybe "badges"? Maybe a "roll of honour"? Even introducing "member stats" so people can see how they stack up against other members across all selected FB projects?


So, I think it is worth carrying on but with perhaps some minor changes to help make it more worthwhile - some ideas have been put forward before but there is no guarantee that any little changes would bring an increase in "team or member involvement" so, any changes would need to be introduced at different times to see if an increase in participation occurs. But that might require "testing" to see if each idea is viable?

I am hopeful that other forum members can put forward their ideas too, so hopefully some renewed interest can take place before next year?

regards
Tim
scole of TSBT
Posts: 31
Joined: Thu Dec 31, 2020 1:53 am

Re: Future of the Formula BOINC

Post by scole of TSBT »

I would like to see Formula BOINC continue. I think the committee was a good idea but being a member prevented one from participating in any planning or strategy. If there was a way for committee members to keep Formula BOINC running and be able to participate with the team that might incent more people to be on the committee. As for project selection, I say unless a project says "no, don't choose us" then let it be an option for a randomly chosen Sprint projects. Sure, some projects will struggle but that's the way it has always been with BOINC comps and we all survived and continued to crunch. As for all the rules suggested, I don't think it's necessary. As a matter of fact it's a waste of time because there will always be a way to skirt the rules.
UBT - Timbo
Site Admin
Posts: 243
Joined: Wed Nov 23, 2016 1:21 am
Contact:

Re: Future of the Formula BOINC

Post by UBT - Timbo »

scole of TSBT wrote: Sat Jul 17, 2021 1:45 pm I think the committee was a good idea but being a member prevented one from participating in any planning or strategy. If there was a way for committee members to keep Formula BOINC running and be able to participate with the team that might incent more people to be on the committee.
Hiya,

There's nothing to stop a committee member from participating, BUT the issue that has cropped up is when "inside information" might be used in advance (by way of bunkering) to "boost" a teams score.

That has now almost been eliminated, as the Sprint Committee votes on suitable projects and then the projects with the most votes go to a "decider" that Sebastien chooses.

In my case, I deliberately didn't crunch any tasks for the first few Sprints...but since most of the UBT members are not taking part this year (as they are taking a sabbatical), so I'm now crunching without any bunkering, as UBT need some points else we might get relegated into League 2 :-( But 2 hosts with 2 older generation NVidia GPU's is hardly going to generate a lot of credits.
scole of TSBT wrote: Sat Jul 17, 2021 1:45 pm As for project selection, I say unless a project says "no, don't choose us" then let it be an option for a randomly chosen Sprint projects. Sure, some projects will struggle but that's the way it has always been with BOINC comps and we all survived and continued to crunch.
I can see where you are coming from on this...but the issue is that some teams did bunker a significant number of tasks in advance of Sprints starting...and then once a Sprint had started, they then re-enabled network access and swamped the upload servers with lots of data. Under normal circumstances, the same amount of data load spread over a few hours or days, would have been fine...but we know that some projects have restricted bandwidth and/or limited CPU capability to deal with thousands of up/downloads at once. :-(

In principle, project admins have been asked, in advance, if they want to be involved in marathons and/or sprints - and this has been fine most of the time.
scole of TSBT wrote: Sat Jul 17, 2021 1:45 pm As for all the rules suggested, I don't think it's necessary. As a matter of fact it's a waste of time because there will always be a way to skirt the rules.
Some rules do make sense...as, IMHO, there does need to be a levelling of the playing field to give more teams a chance....just as in F1, with the new financial "budget caps" being introduced to allow better competition.

The question though, is how far does one have to go, as ideally, keeping it simple, usually allows more people to join in the fun and making up more rules only usually discourages people from getting involved (in my experience).

regards
Tim
Icecold-TAAT
Posts: 50
Joined: Fri Dec 25, 2020 6:18 am

Re: Future of the Formula BOINC

Post by Icecold-TAAT »

I had brought this up before( viewtopic.php?p=1982#p1982 ) but since we're discussing rules - There really needs to be a clearly posted code of conduct. Even if the posted code of conduct was "hey, there's no rules, do whatever you want!" it would at least cut down on one of the issues that has come up which is that each team(informally) has their own values of what they think is acceptable or fair, and then it causing friction when other teams do stuff that goes past that. At least if it was spelled out with a code of conduct what is acceptable and fair(even if people didn't necessarily agree with the exact rules decided) then it would be hard to find fault with "tactics" other teams take as long as it falls within that. With the current state of it being kind of a free for all, but not really a free for all, and there being a somewhat loosely agreed upon set of acceptable tactics leaves a lot of room for friction. It should either be spelled out that there specifically is no rules, do whatever you want and nobody should have hard feelings regardless of what any person or team does, or it should be spelled out that there are specific rules.

As far as the main topic of this thread, I'm sure there are modifications that could make FB better. I appreciate Sebastien being up front and honest about where things stand and to try to figure out a solution rather than just shutting it down without asking for input/help.
User avatar
sebastien
Steering committee
Posts: 112
Joined: Thu Jan 01, 2009 2:59 pm

Re: Future of the Formula BOINC

Post by sebastien »

Thank you for your interest in the FB.
In order to improve the FB, I suggest you use a method similar to starfish.
In a post, could you list the strengths and weaknesses of FB. Also list what it should be created, what it should be changed and what it should be stopped.
Write only one sentence for each point. Try to balance your lists. (Same number of strengths and weaknesses)

For example:

Strong points:
  • FB helps promote small projects with marathons
Weak points:
  • Few teams participate
What to create:
  • Create features (badges, honor roll) to help teams promote FB to their members
What to change:
  • Open FB to other teams through a registration form before the start of the year.
What to stop:
  • Cancellation of sprints because the projects are overloaded.
UBT - Timbo
Site Admin
Posts: 243
Joined: Wed Nov 23, 2016 1:21 am
Contact:

Re: Future of the Formula BOINC

Post by UBT - Timbo »

Icecold-TAAT wrote: Sat Jul 17, 2021 2:57 pm I had brought this up before( viewtopic.php?p=1982#p1982 ) but since we're discussing rules - There really needs to be a clearly posted code of conduct. Even if the posted code of conduct was "hey, there's no rules, do whatever you want!" it would at least cut down on one of the issues that has come up which is that each team(informally) has their own values of what they think is acceptable or fair, and then it causing friction when other teams do stuff that goes past that. At least if it was spelled out with a code of conduct what is acceptable and fair(even if people didn't necessarily agree with the exact rules decided) then it would be hard to find fault with "tactics" other teams take as long as it falls within that. With the current state of it being kind of a free for all, but not really a free for all, and there being a somewhat loosely agreed upon set of acceptable tactics leaves a lot of room for friction. It should either be spelled out that there specifically is no rules, do whatever you want and nobody should have hard feelings regardless of what any person or team does, or it should be spelled out that there are specific rules.
Hiya

I think FB has been running along "informal" lines, and things have been kept simple...but as time has moved on, a few teams have been looking to find ways of ensuring that their team "wins", so "good sportmanship" practice seems to have been cast aside.

Though, given that there's little to actually gain other than a few posts on a few forums, it does seem rather pointless to expend so much time and effort, just to win "nothing". :shock:

OTOH, I think it is imperative that some way of "levelling" the teams, would be worthwhile...which can be done simply, based on tweaking the way the points are awarded:

So, some ideas I came up, which might be considered:

1) All members must register for a specific team for each year, using a specific BOINC CPID - this prevents teams (with lots of members) winning everything, even if most of their members know nothing about FB

2) Teams could be limited to a specific number of members per team - maybe 10 or 20 say?

3) Teams could be allowed to have multiple "sub-teams" so you can have a 1st team (say Taat-A), a 2nd team (Taat-B) and so on...etc. This could allow extra competition between members of the same team (a bit like the competition between drivers within the same team in F1).

4) Teams might be allowed to pick and choose which members they want in each of multiple "sub-teams".

5) Maybe 2 or 3 times per year, the 3 Leagues get "shuffled"...so all 3 leagues are "sorted", with the Top 25 Teams (in terms of either FB points or total BOINC credits) being assigned to League 1, the next 25 into League 2 and the remainder into League 3. This will mean that the most ACTIVE teams will be competing against each other in League 1, while less active teams will then be pitched against other weaker (or inactive) teams in descending order in Leagues 2 and 3.

6) It might be that some members with multiple hosts could set-up multiple BOINC usernames...so, teams could "arrange" for some users to set up new BOINC accounts for a range of hosts, and thereby "fix" how each of the "sub-teams" could earn credits....but if they did that, then their main account would not benefit (from these "other" accounts) as BOINC does not allow merging of accounts set up under different email addresses...so, there's no real point to doing this.

Icecold-TAAT wrote: Sat Jul 17, 2021 2:57 pm As far as the main topic of this thread, I'm sure there are modifications that could make FB better. I appreciate Sebastien being up front and honest about where things stand and to try to figure out a solution rather than just shutting it down without asking for input/help.
Agreed...FB overall is a great opportunity for smaller projects to get some useful research done and for FB members to feel that they are contributing something, along with many others over a given amount of time. And trying to make FB "better" is a step in the right direction...though it is unlikely to please everyone, if any new "rules/code of conduct" make it harder for some people to try and "manipulate" the placings in either the Sprints or the Marathons.

regards
Tim
scole of TSBT
Posts: 31
Joined: Thu Dec 31, 2020 1:53 am

Re: Future of the Formula BOINC

Post by scole of TSBT »

One issue with Formula BOINC, it is a year long commitment. To really compete, a team must focus on the marathons as well as the sprints for the entire year. Perhaps shorten it to a once per year event, like the Pentathlon. I don't follow F1 but it looks like there are 23 races this year? You could have a 23 discrete single day events, all in a row or on the same day as a F1 race. This shortened format might make it more appealing by not requiring a full year of attention to it.
UBT - Timbo
Site Admin
Posts: 243
Joined: Wed Nov 23, 2016 1:21 am
Contact:

Re: Future of the Formula BOINC

Post by UBT - Timbo »

scole of TSBT wrote: Sun Jul 18, 2021 9:30 am One issue with Formula BOINC, it is a year long commitment. To really compete, a team must focus on the marathons as well as the sprints for the entire year. Perhaps shorten it to a once per year event, like the Pentathlon. I don't follow F1 but it looks like there are 23 races this year? You could have a 23 discrete single day events, all in a row or on the same day as a F1 race. This shortened format might make it more appealing by not requiring a full year of attention to it.
Hiya

Originally, as I recall, FB was a Marathon (year-long) event with 124 teams earning FB points, and with just one "league table" (and UBT finished in 12th place overall in that inaugural year :-) ):

https://formula-boinc.org/index.py?year=2007&lang=

...and the Sprints only started in 2017 by which time, there were 3 leagues:

https://formula-boinc.org/index.py?year=2017&lang=

So, ditch the Marathon and just have the Sprints? I think that would be quite likely to put people off from competing, as a few teams do appear to have a large number of members who will compete...but most teams do not....so it'll end up with just the same few teams generating lots of credits every other week (or so). And it'll be quite boring, in the same way as GRC winning most Sprints last year.

(and yes, the Sprints were taking place the same weekend as the F1 races, but the pandemic changed the schedule over the last 2 years :-( )

Taking some ideas from the Pentathlon does make sense - they require team registration (but once a team is registered, individuals credits are included automatically, just like with FB) - but they only choose 5 projects crunching over a 2 week "window".

And FB is trying to do more, over a longer time frame. So, helping maybe 23 projects through the Sprints and a few more via the Marathon.

regards
Tim
[H]pututu
Posts: 5
Joined: Fri Sep 28, 2018 6:07 am

Re: Future of the Formula BOINC

Post by [H]pututu »

...and the Sprints only started in 2017 by which time, there were 3 leagues:

https://formula-boinc.org/index.py?year=2017&lang=
I remember the good old days when the sprints were first introduced and gave small team like us a chance to beat gridcoin back in then in terms of total credit granted. [H] took the top spot in league 3 in 2017 with the total credit of 1252 points which still stand as a record as of today :) . It seems that this ignited some interests in FB. We then took again the second highest credit score of 1234 in league 2 in 2018. It was fun back then and the rest is history....
Post Reply