New Race engine suggestions

[lee1950]

Jan 12, 2012 19:30:41 GMT 1 JoeLag said:

3. See that the new engine is ready for season 3. Do some more balancing to ensure that attacks in the first part of the race have higher chances of success. Hope that it is possible for less competitive teams to snatch some points and price money by concentrating on the earlier intermediate sprints.

Yes. Those early escapes don't have to win very often, but they should be able to get away consistently and sweep up intermediate sprint and KOM points. Most of the time they will be caught by the peloton, which is fine. But atm no one seems able to escape.

Simple thought:

- Re-tune the race engine to make it easier for small groups (1-3?) get away in the first 1-5 parts of a race. Small groups would be 95-100% doomed to be caught, but would get some points. So smaller teams could copy what small RL teams do in grand tours. Their chance to stay away to the finish might be affected by the average Fitness of the peloton, or the average effectiveness of the peloton (Fitness*race effort) so late in the week they might be more successful as riders are tired.

- Make it easier for large groups (10+?) to get away, at any point in the race. Larger groups would be more dangerous, and might get caught less often (90%-100%). Their chance to stay away might be affected by how close the highest placed rider in the break is to the leader of the race. The further back in time the highest placed rider in the break, the higher the chance they will stay away. Also affected by how close to the finish they escape. The nearer to the finish the greater the chance they stay away?


Complex thought:

Maybe we could further define the Helper role, enable managers to select which parts of the course they want their helpers to work, and limit the number of times a Helper can act during a race.

Limit Helpers to 'n' actions when riding at 70% effectiveness: Examples
- 100% Fitness * 70% effort = 70% effectiveness = 7 Actions
- 80% Fitness * 70% effort = 56% = 50% effectiveness = 5 Actions
- 65% Fitness * 95% effort = 61.75% = 60% effectiveness = 6 Actions

Every 10% of effectiveness = 1 Action for a Helper. Ignore fractions.


Give Helpers Actions defined as:

A) Riding to bring back an attacker = 1 Action/each time
B) Helping a flatted or fallen teammate = 1 Action
C) Riding to control the peloton (keeping the pace up so the break does not pull further away) = 1 Action/for every 5 parts of the race
D) Riding to chase down a breakaway = 1 Action/2 parts of the race
E) Assisting with final Sprint = 1 Action

We would need to be able to give orders to Helpers:

a) "Chase down ALL attacks" (will chase down every attack until he runs out of Action points)
b) "Chase down 'N' attacks" (will chase down the first 'n' attacks)
c) "Don't assist teammates" (only way to prevent a Helper with remaining Action points from automatically assisting teammates)
d) "Control peloton pace from part X" (the Helper will begin controlling the peloton pace at part X and continue through part X+4. This means the Helper will start taking pulls at the front to hold the position of the break static. Will be ignored if there is no breakaway when the peloton reaches part X)
e) "Chase down the break at part X" (if there is a breakaway when the Peloton reaches part X, the Helper will go to the front and pull hard to chase the break. He will continue until the break is caught or he runs out of Action points.)


This is still pretty rough, but the idea is to limit how often a peloton would chase down an early break.

Tactically, only teams that want to ensure a sprint finish, or launch a late attack might want to expend helper energy on stopping early attacks, to conserve energy for other Actions.

[NikeBoy]

Great ideas about the new race engine (maybe Schizm can move this last post from lee1950 and my post to the new race engine thread in the suggestion forum, as this has nothing to do with the Promotion issue. ;D

I would like to say 1 thing offtopic, if I may ( ) : we would like to keep the game as simple as possible to play. More tactics, more complex. But certainly something to keep in mind.

[lee1950]

Jan 12, 2012 22:30:17 GMT 1 NikeBoy said:

I would like to say 1 thing offtopic, if I may ( ) : we would like to keep the game as simple as possible to play. More tactics, more complex. But certainly something to keep in mind.

Agree!

That's why, after I wrote that Helper Action thing, I went back and made a simpler suggestion, ;D.

The Helper Action thing started out small - just to help get the early break away, but kept growing. It would have to be greatly simplified, and mostly automated. Maybe keep some of the actions in the race engine, limited by effort level, and managers would only be given the chance to prioritize the various Helper options?

I'm really interested to see what the development team comes up with for the new engine! It has a lot of promise.

[NikeBoy]

I think the reason why break aways are mainly unsuccesfull, is the low sprint and TT skill of the attackers.
These skills are quite important for attackers and no one really trained them yet, so what do you expect.
If a lot of attacks are succesfull when having sprint 2 or 3 and TT 2 or 3, then ALL attacks would be succesfull in case of sprint 7 and TT 7.

So don't panic yet ...

[ElGringo]

Jan 13, 2012 7:06:08 GMT 1 NikeBoy said:

I think the reason why break aways are mainly unsuccesfull, is the low sprint and TT skill of the attackers.
These skills are quite important for attackers and no one really trained them yet, so what do you expect.
If a lot of attacks are succesfull when having sprint 2 or 3 and TT 2 or 3, then ALL attacks would be succesfull in case of sprint 7 and TT 7.

So don't panic yet ...

Nikeboy, I already trie that.

www.peloton-game.com/funraces_results.aspx?id=144

I put my best rider all atacking in the last check point, and they have all good skills on TT and very good flat, but no one could escape.

[JoeLag]

Jan 13, 2012 7:06:08 GMT 1 NikeBoy said:

I think the reason why break aways are mainly unsuccesfull, is the low sprint and TT skill of the attackers.
These skills are quite important for attackers and no one really trained them yet, so what do you expect.
If a lot of attacks are succesfull when having sprint 2 or 3 and TT 2 or 3, then ALL attacks would be succesfull in case of sprint 7 and TT 7.

So don't panic yet ...

Sorry to contradict: I do panic

I really see the need that you tune the engine to make early attacks more successful because that is what in real life happens. An early attack has to have higher chance of success than a late attack. (And I'm speaking only of the success to get away at all. Of course early attacks shouldn't go until the finish so easily)

But as it is implemented now only late attacks have chances for success at all. There has some balancing to be done.

With your argument of low TT and Sprint skills the late attack should fail, too, at the moment.

Look into real life: On a 200 km mountain tour of the TdF there's usually an escape group after several kilometers of the race and none of the classment leaders minds.

But you surely agree that in a 200 km flat race it is nearly impossible for a group to escape the peloton in the last 30km. (Put aside such unusual and out of the normal occasions in which falls and partial blocking of the road cause the peloton to split)

[NikeBoy]

Did I somewhere write that I would NOT finetune the engine anymore ?

The engine is implemented in a way that it should already be harder to get away when approaching the finish.
So I will look into these results.

[ElGringo]

Jan 13, 2012 11:33:38 GMT 1 NikeBoy said:

Did I somewhere write that I would NOT finetune the engine anymore ?

The engine is implemented in a way that it should already be harder to get away when approaching the finish.
So I will look into these results.

Next funny race I will trie to escape at the 1 check point. I will choose my bests rider with TT, flat and sprint to check ;D

[karelv]

I don't think this has been mentioned before but if so, please excuse me

During the last fun race I noticed that after one of my cyclist succesfully attacked, another of my cyclists (who was assigned the duty of helper) "was one of the guys who gets the peloton going". I don't know the code behind this, but if this means that my cyclist is actively helping the peloton to chase the attacker, in this case my own cyclist, I think it should be adjusted: Helpers don't help the peloton when one of your own cyclists is on the attack, as this is what happens irl.

any opinions?

[adipy]

Jan 26, 2012 17:54:28 GMT 1 karelv said:

I don't think this has been mentioned before but if so, please excuse me

During the last fun race I noticed that after one of my cyclist succesfully attacked, another of my cyclists (who was assigned the duty of helper) "was one of the guys who gets the peloton going". I don't know the code behind this, but if this means that my cyclist is actively helping the peloton to chase the attacker, in this case my own cyclist, I think it should be adjusted: Helpers don't help the peloton when one of your own cyclists is on the attack, as this is what happens irl.

any opinions?

It's really hard to keep things easy in the new engine
let's say you have 1 rider with a good General standing in a particular tour, a few helpers and 1 rider who tries to win the stage (by attacking)
your rider attacks, but another rider who is close to your GC guy attacks too.
what would happen in your suggestion is that all your helpers stop making pace and let the attack group go.
So bassicly you'll get a good result with your attacking rider... but you'll lose the GC

You can make a game like this extremely complicated by having alot of tactical choices (if ... happens then ..., else...; at ... km do ...)
but this will make the learning curve steeper, with the result that the game isn't too appealing

[adipy]

Jan 12, 2012 21:34:31 GMT 1 lee1950 said:

Give Helpers Actions defined as:

A) Riding to bring back an attacker = 1 Action/each time
B) Helping a flatted or fallen teammate = 1 Action
C) Riding to control the peloton (keeping the pace up so the break does not pull further away) = 1 Action/for every 5 parts of the race
D) Riding to chase down a breakaway = 1 Action/2 parts of the race
E) Assisting with final Sprint = 1 Action

another possible action: help rider X on Y surface?

[sila]

i don't know why but i hate helpers.
even in real cycling a cyclist performance cant be improve be another. of course that can be situation when a team member can help but very little...

1. Attacker - we already have this option .....just put your cyclist at 100% -
2. Save fitness - already exist - put your cyclist at 70%
3 final sprinter - hire can be made some improvements in RE. In a group the best sprinter should have the most chance of wining.
now i am pretty sure that this doesn't happens.

[adipy]

I don't think that anyone is talking about big multipliers when it comes to helpers, but I do feel that a small bonus would be correct.
a helper in real life really matters! he can get you water, keep you out of the wind, let you drag in his wheel. lead you out in a sprint.
but I think we should limit it. only 1"meesterknecht" for each rider.

[lee1950]

Jan 26, 2012 19:16:53 GMT 1 adipy said:

You can make a game like this extremely complicated by having alot of tactical choices (if ... happens then ..., else...; at ... km do ...)
but this will make the learning curve steeper, with the result that the game isn't too appealing

I like a little extra complexity, but I've also played a cycling game with almost too much.

One way to reduce complexity for the manager is to develop rules for the RE to make it do things that are a bit more RL.

For instance - give the RE logic like this:

"Does lee1950 have a rider in the break?"
- if NO, then Helpers work.
- if YES, then check:

"Does lee1950 have a rider in the main peloton who is in top-3 (or top-5/top-10; whatever is decided by programmers or community) on GC?"
- if YES, then Helpers work.
- if NO, then Helpers do not work.

This would mean your Helpers only chase breaks without a teammate in it, or with a teammate X if teammate Y is sitting high on GC.


This way team managers do not have to make complex decisions - the RE makes it for them. We just have to understand what the RE rules are, and how they work.


If we want more complexity, but still pretty simple, we make the default rule for the RE "All Helpers chase down all attacks" (similar to what it does now??), but two items for the manager to give different instructions. A (1) check box and (2) a drop down box.

1) ( _ ) - "Check here and Helpers won't chase teammates in breaks. The default is Helpers chase all breaks"

2) "Use this dropdown list to select the Attack point you want Helpers to start chasing breaks. The default is attack pt 1"


.....just some thoughts that try to keep the game simple......

[adipy]

Jan 27, 2012 1:25:13 GMT 1 lee1950 said:

If we want more complexity, but still pretty simple, we make the default rule for the RE "All Helpers chase down all attacks" (similar to what it does now??), but two items for the manager to give different instructions. A (1) check box and (2) a drop down box.

I think that having predefined 'default rules' is a good way too keep it simple for some.
If you don't bother with the advanced tactics, you can still just let the race be ridden on the 'standard' settings.
I don't like it when you come into a new game and you have to select 'who will carry water at how many km, how many times for who? how many energy will he put in this effort....' but if this would be implemented with a standard setting (50km, 3 times, all riders,max effort) it would make it a lot more appealing. (stupid example, we don't need water carriers but I couldn't immediately think of a better thing ;D )