[fred]

eyhung, on Mar 7 2006, 08:32 AM, said:

Fred --

I am sorry if my earlier posts came out as critical in tone or content, or did not clearly delineate between statements of opinion and statements of fact. This will teach me not to post replies hurriedly in the morning.

I would be willing to continue this discussion if you wish, either in public or in private. I still believe I have made some valid points which have not been resolved. If you find them (correctly) lacking in hard data, then how about the idea to maximize participation via an initial no-rake policy? That will enable you to collect as much hard data as you can so that you can make an informed decision about the eventual rake. I will of course respect your decision should you prefer to consider this discussion closed.

I would like to reiterate my support for both your service and your committment to innovation, which are still the best in the business. And I remain committed to helping you with constructive and (hopefully) intelligent feedback.

Eugene

I am also sorry if the tone of my post sounded like I was angry with you.

I can assure you that this was not the case. I disagree with some of your opinions, but I know you are only trying to help us make the best possible decisions.

The software has been designed to be extremely flexible as far as rakes are concerned. We can easily change the per hand fee (or change things so that, for example, whover wins each hand is charged a % of his winnings). We can also arrange things so that different Money Bridge rooms have different fee structures.

In my opinion it will be very hard for us to know if our policies in this area are both fair to the players and sufficiently profitable for BBO until players of various levels play a lot of sessions and offer their feedback.

The idea of having a trial period with no rake for the purposes of building up the game has some appeal, but I don't think this will really help us to decide what rake policies are best. Such policies will necessarily depend largely on subjective opinions of the players and they will not really be able to properly develop such opinions if they are playing for free.

In my opinion, if what we are looking for is opinions on various rake policies, the best way to get such opinions is to put such policies in effect and see how people respond after playing under those policies.

I am happy to continue this discussion if you want. Doing it publicly is probably best as I am sure other people will want to contribute as well.

Thanks for your ongoing interest in our site. Once again I apologize if you found my previous post offensive.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[Blofeld]

To those advocating Russian scoring, etc. :

Given that this is all computerised, if you want to score against par it surely makes sense to score versus the par (double dummy) contract on the hand, as this is (presumably) actually findable.

One could then play either total points or IMPs against the par contract. Sure, cards would still be somewhat arbitrary (double dummy analysers never take a 2-way finesse the wrong way, for instance), but there would at least be some balancing measure.

I think best to let GIB declare sometimes, just to give players the chance to relax as dummy one hand in 4. Or possibly give players the option to take over declarership from GIB?

[cherdano]

I don't like IMPing against the double dummy result, because it changes bidding odds substantially:

Assume you can bid a safe vulnerable small slam, or a grand slam on finding the trump queen. Double dummy you can always make this, so the par is 2210. Then the imp odds force you to bid the grand slam, too, since bidding the small slam settles for an 13 IMP loss, while bidding the grand you have a 50 % chance of a push, versus 50% of losing 20 IMPs.

(Russing scoring has some similar issues: Assume your partner opens 1N, 15-17, vulnerable, and you have a distributional hand with lots of spades and 3 hcp. You result will be IMP'ed against, say, -50 (assuming partner has 16 hcp). Just letting partner play 2♠ guarantees a 5-6 IMP gain. Bidding game, you win 12 or lose 2, so you suddenly need higher than 50% odds for your vulnerable game. I think the artefacts are a lot smaller than IMP'ing against the double dummy result, however.)

Arend

[CarlRitner]

It is a different kind of world, to be sure. $200 from Justin spends every bit as sweet as $200 from Jeff Meckstroth. Or as painful in reverse.

Those are examples - you don't know the folks I play with. But their $200 spends every bit as sweet....

Cheers,

Carl
Not that I would ever gamble illegally; that's just a dream I had.

[Blofeld]

I like Cherdano's argument here.

What about total-pointing versus the double-dummy par, though? The odds are then precisely the same as for standard total points, but the variance on cards held, while still present, is lower than normal.

[fred]

I suspect that for the initial release we will use only the kind of scoring that we are using now (with the rake policies still to be determined).

The method has the advantages of both simplicity and that it already works.

I suspect we could introduce any of the other methods of scoring that have been suggested without forcing people to download a new version. Which new scoring methods we introduce in the future will depend, at least in part, on the number of requests we get.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[fred]

Jlall, on Mar 7 2006, 10:48 PM, said:

Fred, I talked to Bob Hamman about the rake, here is what he said (paraphrased)

"A typical rubber bridge club charges 100-200 points for a 4 and a half hour session."

That equates to 2-4 points a board (I'm estimating 50 hands in 4 and a half hours of rubber). He seemed to think 10 points was extremely high and made it unprofitable for all but the very best players.

Thanks Justin, but can you ask Bob if he is sure this is right for right for low stakes games?

I doubt Bob would be much interested in a game for less than 50 cents a point resulting in card fees of $50 to $100 per session. That seems reasonable.

But I can't imagine that clubs would charge as little as $1 or $2 per session to their players who play in the 1 cent games.

I will try to find out what the rates are nowadays at the rubber bridge club in Toronto that I used to play at on occasion. I seem to recall that even several years ago, per session card fees for low stakes games were in the range of $5 to $10. I also seem to recall that the % goes down as the stakes go up (which makes sense to me).

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[FrancesHinden]

A small club down the road from me charged £2.50 for (up to) four hours of 50p/hundred. That includes free coffee, however!

[Gerben42]

That's very good value for money, thinking what a coffee without the bridge costs nowadays

[fred]

We are thinking about changing the rake formula so that the winner of each hand pays BBO a %.

Any suggestions as to what a fair % would be?

A successful backgammon site uses a model like this and charges 10%.

Included in this are transaction fees.

This raises another issue: is it better for us to have a higher % rake and take care of the transaction fees ourselves or have a lower % rake and make the player deal with the transaction fees?

The main advantage of us paying the transaction fees is simplicity.

The main advantage of having the players pay the transaction fees is that players will be able to save themselves money this way. For example, instead of using PayPal to make deposits (which will involve a transaction fee), they could send us a check instead (no transaction fee).

No doubt their are psychological advantages to both methods.

For example, if the customer pays the transaction fee, it may be psychologically bad for some people to desposit (say) $100 and to end up with (say) $97.

On the other hand, other people may find it psychologially bad to pay (say) an 8% rake which includes transaction fees instead of (say) a 5% rake that does not.

Those of you who have argued for a very low rake may not appreciate that these transaction fees represent real costs to someone. Other costs for BBO include the servers that will be required for running many instances of GIB at once as well as the cost of fraud.

Here is an interesting statistic that some of you may find useful in trying to come up with a reasonable formula for the rake:

The average score when a deal is played in the Main Bridge Club is about 330 points.
(uday: the std deviation is about 300, this includes all hands - tourneys, mbc, gibs, 7Nxx-7 to punish P, whatever)

Feedback is welcome.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[asdfg2k]

Just one minor point. For you to suggest that you will accept checks is commendable. I also think it is a disaster waiting to happen. If this thing takes off, which I sincerely hope it does, have you determined how much effort is involved in:

a) opening the mail
b ) determining who the check is for
c) applying it to the account
d) dealing with bounced checks, fraud, etc.
e) other?

I think almost everybody would understand if you have a requirement that all transactions must be handled electronically. If that is limited to paypal in the beginning (or one of the other ecommerce sites that have a good reputation) I know I wouldn't object.

As you know, Fred, you can't please 100% of the people 100% of the time. And to take on an administrative burden that is likely to be overwhelming in order to avoid a percentage of a percentage doesn't seem like a reasonable thing for you to offer, even though I'm sure the offer is being made with the best of intentions.

[Sigi_BC84]

asdfg2k, on Mar 8 2006, 11:31 PM, said:

I think almost everybody would understand if you have a requirement that all transactions must be handled electronically. If that is limited to paypal in the beginning (or one of the other ecommerce sites that have a good reputation) I know I wouldn't object.

NB Paypal refuses to service online gambling sites. I'm quite sure this would also apply to BBO Money Bridge.

--Sigi

[fred]

Sigi_BC84, on Mar 8 2006, 10:44 PM, said:

asdfg2k, on Mar 8 2006, 11:31 PM, said:

I think almost everybody would understand if you have a requirement that all transactions must be handled electronically.  If that is limited to paypal in the beginning (or one of the other ecommerce sites that have a good reputation) I know I wouldn't object.

NB Paypal refuses to service online gambling sites. I'm quite sure this would also apply to BBO Money Bridge.

--Sigi

We asked PayPal and they checked with the lawyers.

The lawyers said "bridge is a game of skill - this is not gambling - we can work with this company".

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[asdfg2k]

As far as the transaction fees go, I think they should be incorporated into the rake, rather than leave the individual responsible. If for no other reasons than economy of scale.

I admit I haven't thought this out completely, but the transaction fees seem like they would be a function of frequency. At one end of the spectrum, every hand generates a transaction. You lose 50 points in a penny game? A $0.50 transaction is generated.

At the other end of the spectrum, I have initially put $10,000.00 into my account and I don't want it touched until the very last moment possible. In this scenario, I play for a solid month (year?) and have a net charge to my account of $1,000. I expect my account to be worth $9,000.00 and that you have processed precisely one transaction against my account. Of course, I can voluntarily chose to close out the account and thereby generate the transaction and along the way I can certainly see what my account is worth (that is, it won't say $10,000 for the entire month - it will tell me where I really stand).

Obviously, to the extent there are per occurence transaction fees, it is in my best interest to see fewer transactions charged. To the extent there are percentage fees involved that do not vary with amount, it is somewhat irrelevant to me how frequently the real charges are generated.

So I guess the question is how much of an internal scoreboard is your computer system going to maintain in order to reduce my transaction costs?

If you come up with a scheme that is super efficient, I dare say that the ultimate transaction fees will be on the order of a percentage of a percentage (3% of 5%), rather than an addition of percentages (3% + 5%).

Once things have been relegated to a percentage of a percentage, however, they become almost inconsequential.

[fred]

asdfg2k, on Mar 8 2006, 10:47 PM, said:

As far as the transaction fees go, I think they should be incorporated into the rake, rather than leave the individual responsible. If for no other reasons than economy of scale.

I admit I haven't thought this out completely, but the transaction fees seem like they would be a function of frequency. At one end of the spectrum, every hand generates a transaction. You lose 50 points in a penny game? A $0.50 transaction is generated.

At the other end of the spectrum, I have initially put $10,000.00 into my account and I don't want it touched until the very last moment possible. In this scenario, I play for a solid month (year?) and have a net charge to my account of $1,000. I expect my account to be worth $9,000.00 and that you have processed precisely one transaction against my account. Of course, I can voluntarily chose to close out the account and thereby generate the transaction and along the way I can certainly see what my account is worth (that is, it won't say $10,000 for the entire month - it will tell me where I really stand).

Obviously, to the extent there are per occurence transaction fees, it is in my best interest to see fewer transactions charged. To the extent there are percentage fees involved that do not vary with amount, it is somewhat irrelevant to me how frequently the real charges are generated.

So I guess the question is how much of an internal scoreboard is your computer system going to maintain in order to reduce my transaction costs?

If you come up with a scheme that is super efficient, I dare say that the ultimate transaction fees will be on the order of a percentage of a percentage (3% of 5%), rather than an addition of percentages (3% + 5%).

Once things have been relegated to a percentage of a percentage, however, they become almost inconsequential.

Transaction fees only occur when you deposit or withdraw money to/from your account. If you do this through PayPal, for example, PayPal charges a fee for the service.

Transaction fees do not apply after every hand.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[Gerben42]

I don't think charging the winner of a hand is right because it changes the dynamics of the game. Suddenly a plus score is worth less. Also if it is a percentage of the score some hands are more expensive than others.

Also this strategy means that less players who start with $100 will go away with more than $100 than the original scheme. That's a bad thing since winners are happy people and they come back.

I would suggest to put it at $0.05 per hand per player and leave it at that for the 1 cent game, and progressively more, but less percentagewise for the higher stakes games, for example $0.30 per hand on a 10 cent game.

[hrothgar]

I have a very strong bias towards "transparency" in business models.
I believe that thing are easiest when revenues and cost structures are aligned.

From the sounds of things, the variable cost of supporting money bridge on BBO is the transaction fee each time players purchase money using Pay Pal or Credit Cards or what have you. Proposed solution: Pass the transaction cost on to the consumer. If it costs you $3.75 for me to purchase $100 of chips, charge me 103.75% for $100 of chips. Please note: If you try to side step this issue by building these costs into the rake or some such you're going to encounter problems if people start making large numbers of small transactions.

In addition to this, BBO wants to make some money. All fine and dandy. Its perfectly reasonable to want a return to compensate you for all the development effort that you are doing. Personally, I strongly prefer a system based on fixed table fees rather than a percentage of the amount of money that changes hand. It doesn't make sense that your revenue stream should be a function of the difference in skill between two random bridge players.

Consider the following: If BBO generated its revenue by taking a percentage of the money that changed hands each deal, BBO would make more money if very good players were matched against very bad players. Conversely, BBO would do worse if two even matched players competed against one another. I'll note in passing that the BBO software also controls pairings: You have the power to determine which players compete against one another. I don't believe for a moment that BBO's management would take advantage of this power. With this said and done, I think that adopting flat table fees could potentially avoid some "conspiracy theories" down the road.

[asdfg2k]

Ahhhh, so transaction charges are really deposit/withdrawal charges.

How about this? Include 1 deposit and 1 withdrawal per month without charge to the player. Anything in excess the player is responsible for. The vast majority of cases will therefore not see any transaction charges. And BBO doesn't end up underwriting Mr. Bunny FlipFlop who decides to make 30 deposits and 30 withdrawals in one month.

If this is not feasible, I see no alternative to having the individual responsible for all transaction fees. I just don't want to see BBO underwriting Mr. FlipFlop's actions.

If it is feasible, I find it hard to believe that the rake would be increased from 5% to 8% merely to cover the transaction fees. More like from 5% to 6%. Or from 5 cents per player per hand to 6 cents per player per hand.

Let's see: I deposit $100 and am given credit for $100. BBO gets charged $3. I play a series of hands at 1cent per point to lose it all (I never win a single hand). The average points scored on a hand is 300 (or thereabouts, from a previous message). So, I am playing roughly 33 hands. If the "normal" rake would be $0.05 per hand, the rake would need to go up by over $0.10 per hand to ensure BBO doesn't lose.

Obviously, people are not going to lose 33 hands in a row, each costing 300 points.

Are there any statistics available that would help determine how fast or slow the average player would win/lose money?

I just did a test on my last 4 weeks and the result was an average of just over 35 points per hand (I won't say whether it was positive or negative ;-)).

At an average of 35 points rather than 300 there are nearly 9 times the hands involved as indicated above, so the rake would need to go up by just over 1 penny per hand to cover the transaction costs.

So I vote to include it in the rake, but not have BBO underwrite those who decide to incur frequent transactions.