John Hurst writes:
Ex World Champion Mike Sovodny once said that the emergence of computer programs such as Snowie had, for him, to some extent lessened the appeal of backgammon. Certainly playing props, at which he was an expert, is now pointless as all you need to do to find the correct move or doubling decision is to stick the position in the computer.
In the past even experienced players would differ wildly in their opinions of the right move and nobody could be sure who was correct; that uncertainty made the game intriguing. Some enthusiasts would spend countless hours rolling out positions to find the right answer and presumably that work would be rewarded by their improved results. Now all a player has to do is ask Snowie or XG. Different players would develop different playing styles, now all players who take the game seriously try to ape the computer as closely as possible.There is no point in trying to develop your own strategies, if your move differs from the computer then it is wrong – it’s as simple as that.
Much the same thing has happened with betting on horses; Internet betting exchanges, notably Betfair, being the equivalent of XG. Betting rings at British racecourses used to be vibrant exciting places full of characters, both bookies and punters, risking their money on their own opinions. Now on-course bookies are little more than commission agents terrified of stepping out of line with the machine. They can’t be blamed for behaving in this way, if they do step out of line they are immediately pounced on by punters who can now find the correct odds simply by checking on their mobiles. Punters may as well sit at home in front of their computer screens. Of course even today’s relatively anaemic betting rings are hugely preferable to the state run pool based monopolies that operate on most racecourse abroad.
As far as backgammon is concerned, as well as their homogenising effects, computers may put off new people taking up the game. Some players may be discouraged by the realisation that the only way they can compete is to practise against computers and then try to copy the computers as closely as possible. New and existing players may be put off by the fact that their every move can be assessed and then they can be informed just how badly they are playing. Previously players could at least have the illusion that they
could play the game to a reasonable standard. No doubt computers have brought some benefits; the level of play is much higher and it is interesting to know the correct play in various positions but perhaps some of the magic of the game has been lost.
Feel strongly about an issue in backgammon? Got something to get off your chest? Write for us! Send contributions to julia@ukbgf.com .
Computers have made the game more accessible to areas where there aren’t any clubs. I enjoy my computer but given the choice of a computer or a human I would go with the latter. One thing a computer can’t do and that’s show emotion, I love looking at a persons facial and body language to offer the cube knowing that it will be rejected so that I gain a point, especially if that person is on the back foot.
Personally, I hanker back to the days when you could play a whole match without someone taking a photo, and didn’t have to listen to people droning on about XG.
It also highlights the differences between the sexes – I’ve never seen a female pause to take a photo during a game.
That said, I’m not exactly ‘with it’, having only recently upgraded from my Nokia brick phone (under duress), am not on facebook, and for whom a tweet is a sound a bird makes – so it’s hardly surprising that I haven’t embraced computers.
For me, I wouldn’t go so far as to say computers have ruined the game, but I certainly wouldn’t lose any sleep if I woke up tomorrow and there was no more XG 🙂
1. For me, it is definitely the precision and excellence of computers that makes backgammon so rewarding. eg in Chess, the computer can only say that one moves leads to a win, whereas another leads to a draw. In backgammon, the computer can give precise measures, of the exact probabilities of game wins and match wins. This delights me.
2. Moreover, I look forward to the perfect computer algorithm, that has cracked all ~ 10^21 backgammon positions, and is absolutely correct.
In Backgammon Funfair, I report slightly fewer positions: 7.41*(10^19). An as yet unknown percentage cannot be reached from the starting position.
While the bots are no doubt superb players, they have not reached perfection – and backgammon is sufficiently complex that they probably never will. It’s quite possible that there are whole classes of positions they misplay (not awfully, just less than optimally), but as a result of their internal algorithms they also play to avoid getting in such positions in the first place. How can we prove it, though? The bots naturally tell us they are playing perfectly in their eyes, and we don’t have a higher authority to go to. And no one person has the time to play enough matches against a bot to demonstrate conclusively that they have an edge.
So the bots can give us a number, an equity difference between two alternative plays. But they can’t explain why the difference arises, what principle underlies the better option – so trying to emulate a bot is a tricky task. We work in very different ways to our silicon adversaries, working with more abstract understanding rather than brute force calculation.
Backgammon played over the board has a multitude of other challenges to add. Can you maintain concentration over a long tournament? In clocked play, can you manage your time effectively? Can you find your opponent’s weaknesses, tempt him into mistakes, use the cube to apply a bit of psychological pressure at the right moment? Can you negotiate hedges well? The bots may have the upper hand with their play, but these subtleties never even arise for them. They may have changed our game, but ruined it? No…
Have Computers Ruined Backgammon ?
For me NO ! In my opinion they have just created more Intrigue , Mystery and Interest in the game !
The more you know, the more you realise you don’t know ! Gaining knowledge from Computers just creates even more questions one wants answered. Such as. 1) Yes that maybe the correct strategy against a player of equal ability, but is it the correct strategy against a player a very poor standard ? A human can think strategy from the moment he or she knows their next opponent then play moves to the strategy they think is what’s needed to win the match. No doubt XG would say Blunder Blunder ” your having a laugh mate ! ”
Also a Computer cannot see the bigger picture !
Example : I was playing a league match in Richard Biddles London League. Free advertising there Richard. I’ll have a Bailey’s with Ice when I see you next ! Anyway I was playing the Top Dog ( yes your right ) He had won the first 11 point match. On the second match after a few games I cubed him to 4 which mystified him ! I said to him “I will explain afterwards”He had to take which he did. Basically I needed a 6 to hit his blot. If I hit, the state of play meant I had a huge percentage of winning a gammon and the match. If I missed he would go on and win the match. So he was favourite to win. I did miss and he did win ! Now for my reasoning : XG would have made my play a Blunder. However as I said it cannot see the bigger picture. I actually needed to win the Head to Head in this match to come Top and go up. Although I did need to play another player I was confident of winning both of them which I did. So for me this was the final. Because of the score, if he won this game I could not win the Head to Head because of the score in the first match. So for me this was 100% the correct play as had I hit I would have won overall and come Top !
Another problem with computers which Mystifies me. Hopefully someone out there can answer this question :
Can XG take into account the importance of the match ?
XG does its stuff over a string a games, I’m guessing, not taking into account the value of the match !
To me matches have different levels of value. Let’s go to the extremes, and everything else falls in the middle. Your local weekly club match which you will no doubt have hundreds of matches over the year. In these matches I’m sure one would want to play as close to XG as possible. Due to them being part of a Long String of Matches.
Now for the other extreme. World Final Monaco 2016 !
I would view this match as a One Off ! This surely cannot be classed as a being part of a normal string of matches which XG needs to Roll Out to ascertain the best Moves and Cube action.
It’s hard for me to explain but I’m sure you get the gist of what I’m saying 🙂
If this is the case, then should one adjust their Play, Cube Action and Strategy to what they feel is needed even though they know XG would play differently?
And also who would be right. XG or the Player ?
For me Humans have created Computer Programs to improve their knowledge. But never forget. At all times we can see the Bigger Picture and hence will always be one step ahead ! Computers do give us useful knowledge and like I said in the first place “The more we know, The more questions we want answered ! ”
Sean Clennell
XG gives the same equity value to winning a match 11-0 or 11-9. A win is a win.
For matches within the London league, where match scores are important, a win isn’t a win, and so we should evaluate ‘league equity’ instead of ‘match equity’ to get the right answer. XG doesn’t do this.
Computer or ‘AI’ opponents are a great way to learn or practise your game. We should also not forget that the digital version of the game may also allow us to play people in different countries and of different skill levels. This alone makes having a digital version of the game worthwhile.
I for one don’t fear the loss of my rating whilst playing…regardless of what the computer says…. I fear the “loss of the game” far more….so if I think I can gain an advantage by sending a cube one roll too early when I am 65% to win but knowing I will be penalised by the analysis….so be it…. I still win 2 out of 3 games…. I know the percentages but those are over a huge number of games rolled out electronically….not over 5/7/9 or 11 points. Maybe that’s why I’m not in the top echelon of UK players….but I also don’t hold any fear of playing anyone….if the dice gods are with you….XG is not going to win or lose you the match ;)))
The good news is that however much we try to imitate the computers, we can’t. Some players get fairly close to it, but even they will fall down in really complex or unusual situations. All players, even the best, are human beings with characteristics (cautious, agressive, flamboyant, steady) and as such will have developed a playing style that is all their own.
The flip side of this is that, as human beings, we can do what the computers can’t do, and that is to play the opponent. XG and Snowie will always play the same, regardless of the opposition, whereas even I will know to adjust my checker play and most definitely my cube decisions according to who I am playing.
My feeling is that we should celebrate the advancement of computer programmes as excellent learning tools, but also rejoice in the fact that we are not machines. However much we try to imitate them, we will continue to have our own strengths and weaknesses, but can still enjoy the game through experience and challenge.