Stormtrack Chase Game

[Skip Talbot]

Thought this needed it's own thread from the future of ST one. We used to have chase targeting game on here, and for years, there was talk of bringing it back. Basically, you're presented with a setup, pick on a map where you'd chase, and then scored on how close you got to the event.

I'm able and willing to bring the game back. I think it'd be fun and a way to promote interest in Stormtrack, forecast discussion, and community. I've got several other projects I have to attend to first before I can consider writing any code for the game, but I thought it'd be a good idea to open up discussion on how it might work. What are your thoughts and suggestions? Is it even a good idea? Some questions:

Real time events or historical?
Competitive scoring or just for fun?
Open to public or closed group?
Moves during game play viewable to public, players only, or nobody (no cheating! or public leaching!)?
Moves after game is over viewable to public or players only?
How many moves?
Scoring based on any LSRs, tornadoes only, surveyed tracks?

Here are my thoughts:
Facebook login, because it sounds like an easy way to handle login and user accounts. I don't have access to Stormtrack forum logins, and interfacing with vBulletin might be a piece of work.

Real time play, the game runs daily and is scored based off what happened that day. Historical events would require a lot more maintenance and admin work as the data has to be fed to the players and each day setup.

Map only. You plot and see moves on a Google Map. All additional data is left to the player to go find. Use your own model sites and observations. Just like real life.

First move is due by 0z the night before the event and your move can be anywhere. Just like in real life, if you're chasing from a distance, you'll have to decide to be there in advance. This should hopefully keep players spread out and nudge people away from playing the day 1 tornado probs or the HRRR plots.

Additional moves due before 12z, 18z, and 21z. These moves have a maximum allowable distance. For example, you're 21z move may be up to 100 miles away from your 18z move. Google Maps will route between your moves to determine the mileage and the game will use this route to determine if the move is acceptable. This will keep players from unrealistically jumping all over the map. Once you commit to that Caprock play, there's no racing up to Grand Island when the warm front starts looking good.

Scoring is based off proximity to LSRs, type of LSR, and number of moves made. Tornadoes are worth the most points. Nail your target with your 0z move? Stay put then, and get a better score, not unlike golf. Hopefully there's some admin tuning of the LSRs as the prelim LSRs may not be very reliable.

Players use play money, or tokens. Every player starts out with a fixed amount of tokens to use for their season. Playing a day uses their tokens. Making moves uses their tokens. This will force players to budget their chasing, just like in real life. If you chase every day in March and April, you may be out of tokens before the end of May. I'm not sure if players should be allowed to earn tokens through targeting successes. Chasing is a hobby or vacation for most of us, not a competitive job, and maybe this game shouldn't promote that kind of ideology.

Daily and seasonal winners on a board.

Ideas? Thoughts?

 

[Jeremy Perez]

Thanks for looking into this, Skip! Taylor Campbell set up a storm chase game last year that really helped me focus on my own crappy style of forecasting. Unfortunately, it didn't encourage discussion, and forecast methods of the different participants weren't shared, just results. Still, I got a lot out of it.

Your suggestion of working with live/current events would definitely make administration/prep much simpler for you and anyone else setting it up—and it would allow for flexibilities for model choices and other current resources participants could use. The 00Z target with smaller moves at day 1 is excellent, although I don't know if the additional moves would create a mess for administration. Point bonuses for staying at the 00Z target vs. moving, as well as a virtual budget of tokens also sounds great.

I could see myself playing Taylor's 2015 game every possible day for the practice, and then budgeting out on the ST version for even greater realism and discussion. Thanks for being willing to get something rolling here!

 

[Patrick Marsh]

Skip, I'm willing to help with writing some of the code here. Just let me know how I can help.

 

[Skip Talbot]

Thanks for the suggestions, Jeremy.

Awesome, Patrick! I'll definitely take you up on that coding offer. You'd probably know best about what sources to use for LSRs, storm data, and parsing it for the game.

Some folks aren't on social media that would like to play. Anybody have any suggestions for alternate methods of handling logins?

 

[Marcus Diaz]

Would it be possible to just get a cheap site and have people create a profile on there to register? There you can have login for people that want to use FB, Twitter, G+, etc...and people that don't have social media (i.e. me)

Take Imgur for example. You just click whichever handle you have associated with your name, and it logs you in automatically:

 

[Rob H]

Skip, you can count on me for coding help.

 

[Danny Neal]

While I know nothing of coding, I can certainly assemble data for the days chosen to run the game

 

[Skip Talbot]

While I know nothing of coding, I can certainly assemble data for the days chosen to run the game

I was thinking the game would run daily, using real world storm reports in real time, not archived or historical events. Running historical cases is interesting and a great test of forecast strategy, but it's also a huge amount of work to gather and deliver all that data. The other problem I have with the historical forecast cases is that the data is never what I actually use in chasing. Letting players go and get there own data from sources that they normally use like COD, TwisterData, Earl, Reed's Facebook page, or the Falmer's Almanac will let the game play out in a more realistic way.

 

[Marcus Diaz]

And although it's not a requirement, we ask that you use your real strategy if you decide to chase in real life. It just really makes this more about nailing a forecast rather than accumulating points.

 

[Rob H]

I was thinking the game would run daily, using real world storm reports in real time, not archived or historical events. Running historical cases is interesting and a great test of forecast strategy, but it's also a huge amount of work to gather and deliver all that data. The other problem I have with the historical forecast cases is that the data is never what I actually use in chasing. Letting players go and get there own data from sources that they normally use like COD, TwisterData, Earl, Reed's Facebook page, or the Falmer's Almanac will let the game play out in a more realistic way.

To expand, I've looked into this in the past. IEM has some great, high-res reflectivity that's already in Mercator projection which jives nicely with Google Maps. I think I figured out that with just reflectivity, warnings, and LSRs, it was like 4-7 megabytes worth of data on an average chase. You have a few dozen people playing, and you do 2-3 games, and you've blown through several gigs worth of data. So having an online solution would be expensive. The alternative is to put it all on a DVD, but then there's a cost involved and the participation goes way down.

Any work I do on a real time chase game, I'd try to do with the possibility of a historical version of the game in mind so code could be re-used, but it would be difficult to offer a historical version for free. Maybe have a one-time download of the data as a torrent (omg a legit use for bittorrent)?

 

[David Wolfson]

With respect to the "armchair chase game/rally/trainer" as with other valuable elements of ST, I think players should be qualified on the basis of their ST membership. The game engine would only be available to someone logged into ST and perhaps some other qualifications associated with their ST membership status. I don't see that a separate authentication process is needed, or that it does much other than to depreciate the unique value of ST membership, but maybe there's a detail or some other considerations I'm not getting. In this context, allowing "outside" access to the game becomes administrative details associated with ST membership that can be tweaked separately.

I'm not much up on interactive web coding, but I should think that, then, a temporary "cookie" registered to a qualified player is enough and avoids the hassles of https or similar cryptographically secure protocols. For any game-related action the game server checks the cookie for authority to perform the action.

The highest level of the action hierarchy is access to a game-home page. What's on that page is the next level of detail up for discussion here. To start that off, personally I'd appreciate having a one-click list of some real-time basic chasing data pages -- SPC mesoscale, COD satellite, RAP surface, etc., etc. Current ST members, e.g. mikegeukes, have a lot of work invested in more comprehensive versions of this already. An easily accessible one-click list of a small subset would be handy for both the game and at times in RL. I'd think that the other main element of the game-home page is a menu of game actions TBD.

 

[Skip Talbot]

I think players should be qualified on the basis of their ST membership. The game engine would only be available to someone logged into ST and perhaps some other qualifications associated with their ST membership status. I don't see that a separate authentication process is needed, or that it does much other than to depreciate the unique value of ST membership, but maybe there's a detail or some other considerations I'm not getting.

Ideally, you'd come to the forum, and there would be a link at the top in the header for the chase game and then you're in the game automatically using the sign in for the forum.

I'm trying to keep this easy from a development standpoint though. I don't know what's involved with tapping into vB's authentication system for your own third party webpages and databases. Using a service that's already setup for this purpose like Facebook or oAuth makes it so much easier. A quick search showed that there might be some third party plugins to allow authentication on other sites using a vB forum, but here's the other problem. I don't have access to the forum's database or software. In an ideal world I would, but realistically speaking here, let's assume that Tim is keeping admin access to this forum private and that the management and activity of this forum is going to stay the same. The only way this chase game is going to get done is if we make it ourselves on a separate site.

An easily accessible one-click list of a small subset would be handy for both the game and at times in RL. I'd think that the other main element of the game-home page is a menu of game actions TBD.

A links section is a great idea and an easy add.

 

[Jeff Duda]

Here's an idea. Might be a way of increasing participation.

Have people pay a small fee to enter the game (say $5 or $10 for the whole season). Then give out a monetary prize for the top three finishers. Obviously the prize pool would depend on the number of entrants, but perhaps you could do something like
1st place: $100
2nd place: $50
3rd place: $25

Obviously issues on internet gambling and other prize money laws would have to be considered. I don't honestly know where this type of scheme would fall regarding the law.

 

[Tom Stolze]

This definitely has my interest, I think concerns would come from more casual chasers/enthusiast that might not have the time or want to put tons of time into playing. For myself, less is more, I found that out quick with fantasy baseball, I enjoy the once a week roster change leagues other than the daily roster changes. Of course, this would not work with weather, but consider those that might not want to spend more than a few minutes each day making a selection.

 

[Rob H]

We're initially thinking "pin the tail on the donkey" type scoring, with some weighting on *objective* events. We can't evaluate quality of a tornado in a consistent way, so seeing any tornado is worth more than seeing large hail, which is worth more than seeing any severe report, which is worth more than seeing blue skies! A limited number of days you can chase in a year and a limited number of moves per day with bonus points for moving less frequently are two of the main ideas we're working with.

Prizes would be cool, but would complicate things quite a bit. Spit out ideas if you have em!