Thursday, July 27, 2017

Solo Race #6... Francorchamps

First Corner at Francorchamps -- Vettel and Stewart
lead the field while Raikkonan and I bide our time.
One of my favorite tracks saw a LOT of carnage on the last lap as the long straights took their toll on the field.  Only half of the cars finished!  Which half?

For this race I did a turn-by-turn Google Photo Album as I experiment with every conceivable method of reporting these races.

This is not too different than how I reported Oyama, but its easier for me to do it in Google Photos and also I planned this from the beginning this time.

Also, Google Photos makes it easy to flip through the pics.  If you click on the (i) icon in the top right, my notes will appear (I added notes to most of the slides but not all).

Lost track of the other races in the series?  Check out the page where I link to all the race reports.

Thursday, July 20, 2017

WBC is Coming and so is the Inaugural Organized Play Champion

Current Organized Play Rankings
The final event for the first CFR Organized Play season is coming.  Although the biggest news may be that I've made a mid-season change to how the rankings are scored.

The In-Person Bonus
I've been noticing this year that online races and especially online tournaments have a scoring advantage over in-person races.  Online tournaments can more easily accommodate larger numbers of participants and ladder-style series concentrate higher ranked players into the same races, making those races more valuable.

I love online tournaments and races.  But I don't want in-person to be devalued.  So, I'm now adding an in-person bonus to the Average Field Rating to create AFR+.  For this season, the in-person bonus is + 0.18 for in-person races and +0.33 for in-person tournaments.

How did I come up with those numbers?  I measured the difference in field ratings between the best and median in-person and online races and series over the last 2 years.  Then I added 0.10 to that difference in favor of in-person.

I've added AFR+ to the race and tournament reports -- which have been updated through PrezCon Summer and sorted by AFR+.  I've also updated everyone's rankings based on this change plus data from PrezCon Summer.

Obviously, this change in scoring will help those who have participated and done well in live races -- mainly the two PrezCons.  Don managed to leap-frog over Fabio into 4th as his win at PrezCon Winter was given more weight.  Brian DeWitt moved into 7th after winning PrezCon Summer.  That said, the three most valuable race wins were online and both online tournaments were worth much more than the two in-person events so far in 2017.

WBC
And that brings us back to WBC.  In the past this has been the largest gathering of CFR players in the world every year.  And it remains the largest in-person gathering although my online tournament has drawn more participants the last couple of iterations.

With the new in-person bonus, ut will be interesting to see what weight WBC ends up with.  If WBC pulls in significantly more people than we've seen in past years, anything could happen.  But as it stands now, Doug Galullo is still the driver with the best shot to un-seat me atop the CFR Organized Play rankings.

More information about CFR at the World Boardgaming Championships.

Tuesday, July 18, 2017

The Tragedy of the Game Commons

I finally got around to reading chapter 3 in Games of Strategy by Dixit, Skeath, and Reiley.  I've fallen way behind Messrs. Aaron, Austin, and Paul.  They recorded their discussion of the chapter and Paul followed up with some written notes as well.

Decision Trees
The decision tree is a method put forward by the authors for how to create a strategy for a sequential move game -- a game where each player takes a turn in sequence as opposed to simultaneously.

These trees or flow charts start with a list of the choices the first player has -- each option being a different branch of the tree.  The second player's options branch off of each possible play by the first player, etc.  So If my choices are A or B and the second player's choices are Y or Z then the tree would look like this:


The four boxes on the right side are the 4 possible end results of this two-turn game.  From here, you can analyze the value of each of those 4 results, then predict the 2nd player's best actions after either A or B, then decide what the first player's best choice is.

Multiple sequential turns can be added on and further branch from here to model the entire game.

Tragedy of the Game Commons
One specific example in the book immediately reminded me of a classic awkward gaming moment.  In the example a handful of neighbors decide, in turn, if they will pitch in to fund the upkeep of a common area.  Although they all want the area maintained, only a certain number of them have to pitch in to make it happen.  After building the decision tree, the game theory "correct" play for the first person is to never fund the upkeep because the other people will.

In economics there is a similar concept called the tragedy of the commons.  Which describes how in real life no one ends up paying to maintain the common area described in the books example.

I think we've all seen this behavior in board games.  One player is winning and something must be done to stop that person.  But stopping the leader does not help the person doing the stopping -- it only hurts the leader.  So no one wants to.

Eventually, everyone teams up to stop the leader or one player is forced to do so and this never feels fun.  In fact, sometimes it feels so un-fun that no one ends up prevents the leader from winning.

Obviously, Tragedies of the Game Commons are something games should strive to avoid.  Can decision trees help us designers see them coming?

Tuesday, July 11, 2017

PrezCon Summer Quick Report


 A couple weeks ago 9 drivers showed up in Charlottesville, VA to contest PrezCon Summer's Championship Formula Racing event.  Don Tatum ran his first event with aplomb.

The two qualifying races featured two older European tracks that I developed recently for the Redscape online series but have never been raced live till now.

Race 1 on Friday saw 8 drivers contest Rouen.  Brian DeWitt won the race with Don and Jim Fleckenstein rounding out the podium.  Tim Mossman finished 4th.

Race 2 Saturday morning found 9 drivers at Castellet.  Jim won this one with Tim and Brian rounding out the podium and Don finishing 4th.

After a lunch break, 7 of the drivers returned for the finals battle.  Clearly Jim, Brian, Tim, and Don had established themselves as the class of the field here.

After 3 laps at fan-favorite Austin Jim, Brian, and Tim were 3-wide and prepared for a dash across the finish line -- shown below on the left.  Brian won that duel and claimed the tournament plaque.

 


Thursday, July 6, 2017

Mexico City solo race is complete

Well, its been complete for a while now but the last of the 4 videos documenting this epic is up on YouTube and all 4 are below.

Mexico City is a track that I have clearly just not figured out yet.  Although my inability to get off the start line cleanly and my busted decel certainly hampered this race.

Stewart lost ground to the competition again though as he used up his wear too early again and fell back into the pack. It was a good race for the Ferrari as Fangio and Vettel both have good races.

Part 1:


Part 2:


Part 3:


Part 4:


New rankings after 5 races -- half-way through the series:

Tuesday, June 27, 2017

Mexico City solo race underway on YouTube

I've posted the first of 4 videos of my solo race at Mexico City.  Keep an eye on the YouTube channel for updates.  I'll blog it again when all the videos are posted -- probably next week some time.

Below is part one, where I try again to push my start speed and yes, I do mis-shift for the 3rd race in a row.



Tuesday, June 20, 2017

The Basic Historical Strategies for CFR

This is the first in a planned 4-part series looking at the Historical Strategies I built for Championship Formula Racing and what I was thinking.  [ Historical Driver rules, video demo ]

Broadly, I tend to group strategies for CFR along a continuum defined by when in a race wear is spent.  On one end is the Front strategy where wear is mostly spent early in a race.  On the other end is the Back strategy where wear is spent late in a race.

A classic Front strategy would see a driver bid high for pole, start off fast, and spend more wear early.  When you see extreme examples of this, you will often see the driver run through half or more of their total wear through the first third of a race and end up with only a wear or 2 before the last third of the race.  A classic Back strategy would see the driver spend no wear at all for the first third of the race.  Spending most of their wear over the last half of the race.

So lets kick off the series by looking at each phase of three strategies that illustrate this continuum: Front A, Back Standard, and Even.

First a quick acknowledgement that these are not exactly the strategies that were published under these names.  Three minor editorial choices I've made since publication: 1) HDs can push start using only 1 green skill (the start speed test table changed a bit right before publication...); 2) I decided that the red skill chip could be used more places in relation to HD die rolls; 3) there were a few tactics that had 6 symbols on them and my BYO tool only allows for 5...  Links to these version in my BYO tool are at the end of this article.

Phase 1



The main way to measure a phase's intent is to look at the first starred and non-stared symbols in each tactic.  This gives you a sense of how much faster the car will likely try to go.  Think about it this way -- a car that wants to spend 1 wear in the upcoming corner will try to go 20 mph faster than a car that does not want to spend any wear, all other factors being equal.

If you then multiply each of those possible results by the odds of them being first choice, I get a net result of how much faster this car wants to go.  It might not be able to go that fast, but that's on the driver not the strategy.  I'm going to call this "net" in tables below.

Note that anything with 3 circles of any kind, should be considered starred because that is how the tracks are designed.  Also note, that if a tactic does not lead with a starred option, I'll count the first un-starred option for that tactics starred value since its more likely to be used.

Not Starred:
Front A: net +29 mph (52.9% +20, 47.2% +40)
Even: net +5 mph (72.3% +0, 27.8% +20)
Back S: net +0 (100% +0)

Starred:
Front A: net +42 mph (52.9% +40, 27.8% +60, 19.4% +20)
Even: net +28 mph (58.4% +20, 41.7% +40)
Back S: net +5 mph (72.3% +0, 27.8% +20)

In the first phase we can also look to see how likely the car is to push it's start speed:
Front A: 92%
Even: 33%
Back S: 28%

Phase 2:


Front A stays faster here.  Note that Even is much faster in this phase if it has more wear than c.  Otherwise, Back S is actually faster under starred conditions.

Not Starred:
Front A net +38
Even (if w >= c) net +20
Even net +6
Back S net 0

Starred:
Front A net +44
Even (if w >= c) +31
Back S net +20
Even net +14

Phase 3:


Even is now the faster strategy in this middle part of the race:

Not Starred:
Even net +34
Front A net +15
Back S net +11

Starred: 
Even net +34
Front A net +23
Back S net +26

Phase 4:


Lots of conditions here.  But in the end each strategy ends up being as fast as their remaining wear let them.  Without stars, if wear <= c their nets are between 17 and 20.  When wear >= c their nets rise to 20 to 32.  When wear >= 2c their nets are all +40.

But since Front A and Even have been faster earlier in the race, Back S is likely to be faster here with more wear to burn.


Strategy Trends

Over the course of a race, HDs will spend most of their time in phases 2-4.  So let's quick see each strategy's progression over time:

(Not Starred) phase 2 >> phase 3 >> phase 4
Front A net +38 >> +15 >> +17
Even net +6 >> +34 >> +20
Back S net +0 >> +11 >> +20

Some caveats... Front A cars often do not spend much time in phase 4 because Front A HDs often run out of wear early and skip out of phase 4 into phase 5 after a corner or 2.

Even HDs will likely be faster in phase 2 than net +6 because they will frequently have more wear than c and use net +20 options.

Back S' real speed will vary on how much it can save wear in the early race.  Save enough and it will skip into phase 4 early and spend at its higher rates of +32 or +40 for a while.

Hopefully that all made some sense and sounded like real CFR strategies at work.  Next up I'll discuss the variations on the Front A strategy.  Until then, feel free to tinker in the BYO tools with your own strategies.