Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Radio Controlled]

I'm looking for a Bookie to handle the money and betting in DMR. They'd have control over setting the bets and get a cut of the cash, for use with whatever they want. Maybe other things, if they should need or want it. Preferably someone with an adequate grasp on statistics and betting.

Again, I'd love to, but no time. I could suggest an easy way of making the betting more 'realistic' though:

Say there's two peeps, x and y, in a fight to the death. 1 DD gets bet on x to win, and 4 DD on y (two peeps betting 2 DD each) (cause x is missing a limb or so). Starting payouts is bet*2.

Now, the payouts get a modifier which increases as the 'opposition' to your get increases. The modifier for x is 4/1=4, and for y it's 1/4=0.25
This modifier is added to the standard modifier for that type of bet.
So now, if x wins, the payout for his bet is: bet*(2+4) = bet*6 = 1*6 = 6 DD paid, so 5 DD gain
And if y wins, the payout for a single bet is: bet*(2+0.25) = bet*2.25 = 2*2.25 = 4.5 = 4 D paid, so 2 DD gain for each individual better, so 4 DD payed out total that round.
Payouts are rounded down (or up, whatever) to nearest integer.
In case of no people betting on the other person, no modifiers are added.
In case of more than 2 possible peepes to bet on, for an individual contestant his total is put against the total DD of all other bets of all other contestants (in that category).

This way, betting on an unpopular/unlikely to win contestant has bigger potential payouts, but betting on the one almost guaranteed to win doesn't pay as much, but is a 'safer' option.

Thoughts?

[Kriellya]

Well, what happens in ship Alpha the next turn will reveal things about the intelligence and luck of many people...

Yeaaaaah. So, has that message been played yet? Or is that the sod's action next turn?


Cause let's see what our options are for a result here...

Option 1: Order 118 is 'Controllers compromised, kill anyone approaching the bridge'
- This is probably the most likely...
     Potential Solutions: Don't give the order, attack from 'surprise'
                                Give the order, attack when they move (which would be after Flint and Thad have been shot)

Option 2: Order 118 is a valid stand down order, but not a surrender order
- The three of you head past the sods, I wait in the elevator because they would probably shoot me if it's just a 'let them pass' order of some variety.

Option 3: Order 118 is a 'complete surrender' order (best case scenario)
- We walk onto the bridge and take the ship XD

[Parisbre56]

I usually operate on the assumption that people want to live, so they probably wouldn't give an order that will get team AD in immediate trouble, because then they'd be killed.
However, the UWM seems to frequently hire suicidal people (at least from Flint's point of view), so...

We can always take a third option (or actually 3,  now that I think about it). Again, don't want to say so I can see what other people come up with.

[Kriellya]

I'm looking for a Bookie to handle the money and betting in DMR. They'd have control over setting the bets and get a cut of the cash, for use with whatever they want. Maybe other things, if they should need or want it. Preferably someone with an adequate grasp on statistics and betting.

Again, I'd love to, but no time. I could suggest an easy way of making the betting more 'realistic' though:

Say there's two peeps, x and y, in a fight to the death. 1 DD gets bet on x to win, and 4 DD on y (two peeps betting 2 DD each) (cause x is missing a limb or so). Starting payouts is bet*2.

Now, the payouts get a modifier which increases as the 'opposition' to your get increases. The modifier for x is 4/1=4, and for y it's 1/4=0.25
This modifier is added to the standard modifier for that type of bet.
So now, if x wins, the payout for his bet is: bet*(2+4) = bet*6 = 1*6 = 6 DD paid, so 5 DD gain
And if y wins, the payout for a single bet is: bet*(2+0.25) = bet*2.25 = 2*2.25 = 4.5 = 4 D paid, so 2 DD gain for each individual better, so 4 DD payed out total that round.
Payouts are rounded down (or up, whatever) to nearest integer.
In case of no people betting on the other person, no modifiers are added.
In case of more than 2 possible peepes to bet on, for an individual contestant his total is put against the total DD of all other bets of all other contestants.

This way, betting on an unpopular/unlikely to win contestant has bigger potential payouts, but betting on the one almost guaranteed to win doesn't pay as much, but is a 'safer' option.

Thoughts?

Sort of 'auto-balancing betting'. Hmm. Could work, but I think I can game it in this scenario.

So, let's setup the same initial scenario: X has a bet of 1 to win, Y has a bet of 4 to win. Now let's consider a new better (call him B), who bets 1 on each.
(If this sort of betting is not allowed, fine, but I think we'd prefer a system where we don't have to restrict bets like this because betting like this would be a no-win scenario.

X now has 2 to win, and Y has 5 to win. So the modifier for X is now 5/2 = 2.5, and the modifier for Y is now 2/5 = 0.4.
If X wins, B gets 1*(2+2.5) = 4.5 as a reward. A solid win for him. If *Y* wins, B still gets 1*(2+0.25) = 2.25 as a reward. So B can make the risky bet on X, but will never lose money because he will make it back on Y anyway.

I do know there is a configuration of this sort of betting that works, though. It's how a lot of high-volume bets work, because in practice the bookies have as little idea about the odds as the betters do, so they simply use the bets that come in to balance the odds (and, of course, the algorithm they use to do this results in a negative-sum game for the betters, where they can never win more than the bookie receives in bets)


How about reward = b * (1 + base * mult)? Base is the 'base odds', multiplier is as you described it, RC?
So in the original scenario, the base would be 1, such that with an even multiplier, the reward would be 2. The reward for winning X would be b*5, and the reward for winning Y would be b*1.25.

Then in my scenario, B would get 1+1*2.5 = 3.5 for X winning, and 1+1*0.4 = 1.4 for Y winning, so betting on both is no longer a guaranteed gain (and probably not worth it)


Sounds like a solid plan for the team.v.team bids and 1v1 bids, but I think we need more for the survival/death and the winner bids

[Kriellya]

I usually operate on the assumption that people want to live, so they probably wouldn't give an order that will get team AD in immediate trouble, because then they'd be killed.
However, the UWM seems to frequently hire suicidal people (at least from Flint's point of view), so...

We can always take a third option (or actually 2,  now that I think about it). Again, don't want to say so I can see what other people come up with.

Yeah, the options 2 and 3 are sort of equivalent, in that they both result in the sods not being a problem. Though I can now think of a 4th option, which amounts to 'get them to reveal themselves entirely, then kill them'

But yeah, it does depend on how much the controller's value their lives.

I think I'm going to go with what is always a solid default: Hope for the best, plan for the worst. The question is what can you and renegade do to improve your odds of surviving the worst?

[Radio Controlled]

How about reward = b * (1 + base * mult)? Base is the 'base odds', multiplier is as you described it, RC?
So in the original scenario, the base would be 1, such that with an even multiplier, the reward would be 2. The reward for winning X would be b*5, and the reward for winning Y would be b*1.25.

Then in my scenario, B would get 1+1*2.5 = 3.5 for X winning, and 1+1*0.4 = 1.4 for Y winning, so betting on both is no longer a guaranteed gain (and probably not worth it)

Sure, that works too, probably better even. It's only a small amount of extra work, though in your system I'm not sure if rounding up or down would work better. I'd say down, but I'd have to work out a few examples to get a feel for how that works out, and I should already be working on that paper really. And you would have to define a few different base odds (so that bets where there are more than 2 options have higher base odds) for different kinds of bets I think. e.g. base of 2 for whole team, base of 5 for dude dying. Also, if there are no counter bets, I take it your system defaults to base pay (because b*(1+base*0)=b) ? Or do you simply give those people their DD back, no gain or loss?

You could always do away with the basic '+1', meaning it's just base*modifier but then the base odds need to be high enough so that, even if 20 DD are bet on 1 dude and only 1 DD on the other (meaning 1/20=0.05 modifier), those 20 can still earn at least a little (or else they can end up losing money despite betting right). Or maybe be cruel and have those chumps break even, even when betting right, hah. Depends on how much of a mean mofo you wanna be  In real life, there isn't an infinite pool of money to divide from, so the people betting on the guy guaranteed to win wouldn't get much. In case of no counter bets, modifier would default to 0 (or a really low number) cause if nobody is betting against, then what's there to earn for the bookie right?

If the bookie could set up a small computer program to do the numbercrunching for him, it'll just be a bit of initial work, and then just punching in numbers. Hell, one could do it in excel without much hassle.

[Parisbre56]

piecewise could give the bookie an amount of DD at the start of each episode (or better yet every time someone requests DD), about the same (or maybe a bit more so that people can earn money) as that given to people requesting DDs, so then the bookie would have an incentive to make sure he doesn't give away too much money. And  at the end of the episode there could be some standard tax on the earning to ensure the bookie doesn't get filthy rich (which leads to the "possible to give players too much money" problem again).

EDIT: And the bookie can keep the amount of money he has secret so that people can't know if he's hoarding money

[Radio Controlled]

piecewise could give the bookie an amount of DD at the start of each episode (or better yet every time someone requests DD), about the same (or maybe a bit more so that people can earn money) as that given to people requesting DDs, so then the bookie would have an incentive to make sure he doesn't give away too much money. And  at the end of the episode there could be some standard tax on the earning to ensure the bookie doesn't get filthy rich (which leads to the "possible to give players too much money" problem again).

EDIT: And the bookie can keep the amount of money he has secret so that people can't know if he's hoarding money

Que the bookie rigging every bet he can and earning all the DD. ALL OF IT!

[Taricus]

Pretty much my reasoning for wanting to be the bookie. Though I wouldn't be able to do the actual bookkeeping and/or probability adjustments.

Would also make it know that you go along with whom the bookie wants to win too. Admittedly, it's not like a consortium can't rig the games as well.

[syvarris]

Also, I as well have a massive amount of Tinker projects.  Several of which are to clear up old arguments ("Is a max level decomp really three decomps of the previous type?"), and most of which have to do with automanips.  I'm sure that if I get enough data on them, I can break the game.))

"Because the last one  is just three of the ones before it stuck together. It's actually under-priced."

That's from the AM.  I suppose she could be lying, but that's convincing enough for me.

*eyetwitch*

It's convincing enough for me, too.  But somehow it isn't for others.  Seriously, go check the wiki's article on decompensators.  Specifically, the comments.



IN OTHER NEWS, I volunteer to be the bookie.  While I'm not particularly good at statistics, I am a stubborn mule who is good at figuring things out. >.>  And I'm also proficient enough with python to make number-crunching programs.

Although, I'm not entirely certain whether you want the bookie to actually ensure a profit for DMR, or if you just want someone who ensures there's no completely safe bets.  The former is what an actual bookie does, whereas the latter is what caused you to ask for one.

[Kriellya]

I think (PW will have to confirm ) we have two objectives for the bookie

1) Adjust the odds as the conditions of the players change. So, taking the current episode as an example, a bet on Cromwell surviving should have a better payout than standard, while a bet on him dying should have a reduced payout. The system me and RC were talking about was also a potential solution to this, though it will likely need some tweaking (it only really works correctly with either-or bets, like teams)

2) There should be no guaranteed win. By which we mean there should be no set of bets that is guaranteed to pay-out.

[Parisbre56]

Also, I as well have a massive amount of Tinker projects.  Several of which are to clear up old arguments ("Is a max level decomp really three decomps of the previous type?"), and most of which have to do with automanips.  I'm sure that if I get enough data on them, I can break the game.))

"Because the last one  is just three of the ones before it stuck together. It's actually under-priced."

That's from the AM.  I suppose she could be lying, but that's convincing enough for me.

*eyetwitch*

It's convincing enough for me, too.  But somehow it isn't for others.  Seriously, go check the wiki's article on decompensators.  Specifically, the comments.

Note that I did not say that it is not 3 decompensators of the previous type. I said that game-mechanics-wise it doesn't act like three decompensators of the previous type. Which is true, since it doesn't give you three rolls for it to work (it works automatically) and it doesn't take three mind-slots (it takes only one). So it's better to clarify that in the description so that we don't get people confused and asking questions about it. How one clarifies that is irrelevant.

[kj1225]

Sorry if this sounds stupid... I am okay to start posting right?

[NAV]

In the on ship thread.

[TCM]

Is the AM's "E.R. Dating Simulator" Program on the VR System still active? That would be something fun for anyone on the ship with nothing to do.

And guys, when GWG comes back, we really shouldn't keep goading him onto his soapbox again. This goes for everyone, especially a few individuals, including PW and myself. If he ever starts getting out of hand, don't reply until he posts that's he's calmed down or something. If he doesn't, it's no use trying to stay in the ring with a raging bull. I'm not just saying this for his sake, but all of y'all as well, because Toady knows what's up and if you instigate him there's a fair chance your ass is getting in trouble too.