Speedway Promoter Game?

[SCB 0]

Back in the late 1990's through to the early 2000's I ran a free email based Speedway manager game (it was called the ISL or Internet Speedway League and I think at least 2 of the posters here were Team managers).

 

It used Kevin's simulator to run the meetings along with a squad system for managers to select for their team for each round. The reason for mentioning this was that I ran it for between 3 & 5 seasons (I forget which and when I handed it on I didn't keep a copy) and found an irritating problem with the simulation. What it was was that as riders started each season with their finishing average from the previous season (various factors not withstanding such as a minimum number of rides/matches and so on) I found that a riders average tended to drop or rise towards an average of 6. A rider with a season start average of 10 usually lost between 1.5 and 2.5 points off that average by the end of the season whilst a rider on a start of say 3 would end up on 1 to 2 up. The game even had riders with a start of average of 12 but who never recorded a maximum despite riding in most of their teams meetings (in fact any form of maximum was a rare thing, I should mention that the game generally used the 13 heat format from the 70's and 80's). The upshot was that I, and several of the team managers in the game, felt that the simulator was unrealistic with too many 'fluke' results and not enough 'predictability'. I mention this because you may want to review the way the heat results are calculated in Kevin's simulator - sorry Kevin this is not a critism and I never mentioned it at the time (we did have some correspondence in that I sent you the team line-ups for the '65 and '66 BL seasons) as I only realised this issue towards the end of my time running the game and thought my recollection might be of use to Paul.

This is a damn hard issue to avoid! You either find riders trend towards 6 or towards 12/0. It takes ages to find value that work.

[Guest compost]

 

Useful to know. There was an issue in earlier versions (pre-Windows) whereby the results seemed to be too predictable, so I tweaked the algorithm to improve the unpredictability although I may have overdone it.

 

My observation though was actually the opposite to you though - I still thought the top riders won slightly too much, and the 3-pointers lost slightly too much. I also thought the number of maximums seemed about right, but I can't say I did any empirical analysis on that factor. You probably have more practical real-life experience of running the programme though, so I certainly don't dispute your observations.

 

My suspicion though, is the average is the limiting factor, and however you tweak things you're going to see a drift towards the median. The fact the algorithm works on probability means that over the long-term it's never going to be possible for a rider achieve a 12-point average, and by definition that will mean the average of other riders will drift up. I guess other metrics need to be introduced to improve this aspect, but I didn't have 20-odd years of data when I first wrote the program.

 

Hi Kevin

Thanks for not telling me off as I really wasn't having any sort of go.

 

The comments in my previous post were what I found at the time - that the drift downwards in average was always higher than any improvement upwards. And there were very few maximums (paid, full or of the 'reserve' kind - the last being exceedingly rare). I don't remember the versions of the game used (initially it was the DOS version though). One other factor impacting on averages was that managers (in later seasons) were able to make a limited number of transfers into their squad (obviously riders would drop out) so in effect it wasn't a closed system. The fact that squads were used (12 mainly) would also skew final averages as a rider would not be 'guaranteed' a complete season. And rider injuries, generated through your simulator, could be anything between just the match injured in to being out for the entire season.

 

My comment on too many surprise results was based on roughly a 22 match league season plus divisional ko cup and inter-divisional cup (at one point there were 4 divisions operating so there was lot of sample data available).

 

Brings back memories.

[MattK 3,447]

This is a damn hard issue to avoid! You either find riders trend towards 6 or towards 12/0. It takes ages to find value that work.

 

How about riders having attributes each with a score. The value of the attributes (multiplied by a random number) determines who wins a race, rather than averages?

 

Attributes also allow a manager to "train" riders, therefore creating a marketplace to buy and sell. The effects of training deteriorates with a rider's age until the point when it has no effect and a riders attributes start to fall.

[Kevin Meynell 2]

How about riders having attributes each with a score. The value of the attributes (multiplied by a random number) determines who wins a race, rather than averages?

 

It would not be difficult to formulate some key attributes, although the relative weighting might take a bit of thought. The limitation is that many of these attributes other than averages are going to be subjective, and of course there's no ready source for these. Averages are easily obtained and (at least until recently in the Elite League) a fairly accurate measure of rider ability.

 

Of course, you might also want to have track attributes as well, which would have an impact on results.

[BWitcher 12,453]

Back when I was a young teenager I would spend hours and hours devising such attributes to be used alongside the Speedway Scene II board game.

 

Every rider would have ratings for gating, a rating when leading a race, when behind, inside ability, outside ability, adjustments for conditions and also the individual track. I even had a system for riders morale, their experience (more laps on a particular track the better they would become) and even down to mechanics and engine tuning which could give a boost to performance (or not on occasion). All of this was done with pen and paper for the most part so a computer could handle it all a lot easier.

 

I had more fun in putting all of those factors together than actually playing through the season.

 

I converted much of it into a very basic game programmed in basic on the good old 48K Spectrum.

 

Essentially to me, the actual running of the meeting is the easy bit. It's all the rest I have no clue about programming wise.

[steve roberts 9,242]

Back when I was a young teenager I would spend hours and hours devising such attributes to be used alongside the Speedway Scene II board game.

 

Every rider would have ratings for gating, a rating when leading a race, when behind, inside ability, outside ability, adjustments for conditions and also the individual track. I even had a system for riders morale, their experience (more laps on a particular track the better they would become) and even down to mechanics and engine tuning which could give a boost to performance (or not on occasion). All of this was done with pen and paper for the most part so a computer could handle it all a lot easier.

 

I had more fun in putting all of those factors together than actually playing through the season.

 

I converted much of it into a very basic game programmed in basic on the good old 48K Spectrum.

 

Essentially to me, the actual running of the meeting is the easy bit. It's all the rest I have no clue about programming wise.

I, too, had the Speedway Scene Board Game and built a track and adapted it so that I could use the Britain's speedway riders incorporating some of the rules.

[MattK 3,447]

Back when I was a young teenager I would spend hours and hours devising such attributes to be used alongside the Speedway Scene II board game.

 

Every rider would have ratings for gating, a rating when leading a race, when behind, inside ability, outside ability, adjustments for conditions and also the individual track. I even had a system for riders morale, their experience (more laps on a particular track the better they would become) and even down to mechanics and engine tuning which could give a boost to performance (or not on occasion). All of this was done with pen and paper for the most part so a computer could handle it all a lot easier.

 

I had more fun in putting all of those factors together than actually playing through the season.

 

I converted much of it into a very basic game programmed in basic on the good old 48K Spectrum.

 

Essentially to me, the actual running of the meeting is the easy bit. It's all the rest I have no clue about programming wise.

 

Precisely. The mechanics of a race are actually quite simple to model. I put together something simple in a hour using Excel with each rider having six attributes:

Gating

Passing

Speed

Team riding

Reliability

Aggression

The main challenge is getting the weighting of the attributes correct and how much randomness to introduce so that results are predictable, but not obvious.

[AndyJ 95]

Precisely. The mechanics of a race are actually quite simple to model. I put together something simple in a hour using Excel with each rider having six attributes:

 

Gating

Passing

Speed

Team riding

Reliability

Aggression

 

The main challenge is getting the weighting of the attributes correct and how much randomness to introduce so that results are predictable, but not obvious.

Don't forget 'confidence'; a roller coaster of an attribute that plays a huge part. Also perhaps some attributes linked to track type (some excel on the big ones yet struggle on the tighter circuits) and preparation (slick/grippy)

[MattK 3,447]

Don't forget 'confidence'; a roller coaster of an attribute that plays a huge part. Also perhaps some attributes linked to track type (some excel on the big ones yet struggle on the tighter circuits) and preparation (slick/grippy)

 

I've been playing around with Grip. The problem is it's the same for all riders, so you need to come up with some kind of weighting. I'm still trying to get my head around that on paper!

 

Track size is an interesting one, I hadn't thought of that.

 

Likewise Confidence (or I called it Form!) is an arbitrary figure, but I haven't worked out how this effects for outcomes (yet).

[Danny Connor 351]

Sporting dreams does a good game. Expensive to some but riders have stats and attributes other than just averages. Has track sizes and conditions as well as gating passing and track stats. Also has confidence and morale ratings. Fitness too and finances with depth. League, cup, 4tt , pairs. Grand prix. Works with a squad system too

 

If this was a computer game it would de as addictive as heroin

[MassMarauder 22]

Because I have been messing around with Flying Shale and Speedway Mtg  (Shale works on Windows 10) - I was wondering if there was any change on this discussion, anything happening behind the scenes.
I am fascinated with the discussion. I still have a copy of Speedway Scene II on Tape cassette that I used to load onto my Apple II back in the mid 80's along with the board and dice.

I started playing baseball Sims on the PC back in 1986 (when I moved to the USA)  This game was available on PC and Amiga and included separate floppy discs with programs enabling you to edit player attributes, add new players etc.  The developer of this baseball game (for EA) left and took his game engine modded it and uploaded a new version on the web as freeware.  Aside from the regular baseball stats, Players also had 3 attributes that affected their performance, which we are editable - although the loaded stats were based on real performance stats (running speed, fielding %, arm strength) and pitchers had editable stats also (pitching speed and Endurance) .  So in speedway terms, for example, that could be frequency of Tape exclusions (as % of starts), % falls per starts and % engine failures per starts, all of which we know from available stats.  I always thought that one determining factor in a truly great speedway sim would be that the rider database would include lifetime real averages by track(s). We all know certain riders perform miserably at certain tracks and score maximums at others. There used to be a web site where you could see all current rider averages rides etc. by track, by year etc. etc.  I think they stopped updating in 2012?   Of course this would require special algorithm calcs I am guessing?

Anyway - just wanted to pop this back up top to see if it was going anywhere?

 

Edited April 9, 2020 by MassMarauder

[Hackett 220]

Considering the loyal following that speedway has and the tendency of the supporters to be keen on stats and statistics I am surprised a game like football manager has never really taken off. I appreciate that the sport will never have the following of Moto GP or FIFA on various consoles but I believe a Speedway game in the Football Manager format could of worked and be well received. The mobile game Speedway Challenge has a lot of the graphics and ideas right.  Purchase bikes, equipment, vans etc but if you could train and develop the riders that would be an excellent improvement. 

[MassMarauder 22]

Man....I wish there were a way in Flying Shale to control ALL TEAMS.  Have got it going with the 2020 lineups using 2019 averages and assessed for Crump and Pedersen :-)

I have Speedway promoter 2002 on floppy - will install this as soon as I get my floppy drive hooked up later this weekend.

 

[Humphrey Appleby 13,942]

On 4/12/2020 at 1:23 AM, Hackett said:

Considering the loyal following that speedway has and the tendency of the supporters to be keen on stats and statistics I am surprised a game like football manager has never really taken off. I appreciate that the sport will never have the following of Moto GP or FIFA on various consoles but I believe a Speedway game in the Football Manager format could of worked and be well received. The mobile game Speedway Challenge has a lot of the graphics and ideas right.  Purchase bikes, equipment, vans etc but if you could train and develop the riders that would be an excellent improvement. 

Just not a big enough market to justify the investment needed to make a professional quality game these days. Would have to be a real labour of love, and even then I suspect licensing the team and rider names etc.. which would needed for it to stand any chance of success, would rule it out.

Whilst speedway does come with ready made stats and somewhat lends itself to this genre, in fact the multiple and ever changing formats are a bit of a nightmare to deal with. Football is pretty much 11-men-against-11 at whatever level and whatever country, and makes for an easier starting point. Of course it could be done if the will were there, but the fanbase and visibility of the sport just isn't there to justify the end product unfortunately. 

[Danny Connor 351]

On 4/11/2020 at 7:13 PM, MassMarauder said:

Man....I wish there were a way in Flying Shale to control ALL TEAMS.  Have got it going with the 2020 lineups using 2019 averages and assessed for Crump and Pedersen :-)

I have Speedway promoter 2002 on floppy - will install this as soon as I get my floppy drive hooked up later this weekend.

 

If there is any chance of copying speedway promoter 2002 please contact me! Mine corrupted and I've been looking for it for a few years now!