|
|
| From | Message | ||
|---|---|---|---|
|
 Personal Game DatabaseI have been trying to learn from my Database. For example in which opening in the weakest, or at which point in the opening i Mistake. I realized that very often a move shows to be of negative result for me but if i click on it the first few possible move over it are good for me. I am confused I searched for an explanation of how does it work but i cannot get it |
||
|
Well, that might explain it, except that sometimes, as you’ve noticed, the only move option (or popular option) available suddenly turns the percentages on their head, in which case I’m pretty baffled myself. Best only use the database as a guide, I’d say. |
||
|
sure |
||
|
It might help... The results graph and win-loss-draw numbers (I would prefer W-D-L, but never mind) really signify results, and not necessary the objective value of the line. More to the point is the average rating of the players who adopted the line. Look at the size of the numbers. too. To take a simple example, the French Defence (1.e4 e6)- third most popular opening at Master level (average rating over 2300, which argues its soundness. Results: 39% White wins 31 % Draw 28% Black wins. Reduce the 106,000 games to ten. For every ten games, White wins 4, 3 are draws, Black wins 3. That's good enough for me, Take the thing a little farther, and imaginne you're playing with the Black pieces : 1.e4 e6 2.d4 d5 3.Nc3 ... You want to choose between 3...Nf6, 3...Bb4, 3...dxe4 These are the three most popular options. But the pawn exchange is not only a distant third, its results aren't very encouraging either, White winning close to half the games, Black barely one in 6. You might still want to explore where the game might go, and there is every reason to search deeper into this line. But you might decide that exchanging on e4 gives White too much in the centre (In fact, there was a time when such a move was dissed as 'surrendering the centre. Balderdash, of course, but you might decide even a temporary small concession isn't worth the candle). Of the other two, 3...Bb4 is considerably more popular than 3...Nf6, but look at the results White does better against the bishop move than against the knight move - not by a huge margin, but ... it's there. There's a slightly higher average rating for the knight move as well. What should you make of this? I could say 'nothing', but the questions are worth asking, and worth an answer. The difference in average rating between 2317 and 2324 - just 7 ELO points is trivial: that you can ignore. When the difference is approaching 50, then it might be worth taking into account (which means there isn't much in the 2297 for the ...dxe4 option, neither. You could go with the knight move over the bishop, but I would suggest that the differences, though favouring the knight move slightly, aren't enough to make a final decision. There is a good deal to be said for 'letting your heart decide'. Suppose you play a few games. You discover that tou p[refer the types of game that emerge from the one move over the types from the other. I'm a big fan of the Bishop move (the Winawer), myself - I don't think I've ever played the knight move. I just find the types of games that ensue exciting to play. Perhaps it is a little strange to say that the Winawer is my favourite line AGAINST the French, as well! (On my database, my record with the Black pieces isn't that flash, but I did persist in a line that involved retreating the bishop to a5, instead of exchanging on c3. My personal database shows 3 wins, 5 losses, no draws. But further examination will show the opposition in those games has been pretty strong! My record with the White pieces has been pretty good - no losses. Perhaps this conversation can be extended if you mention which opening - or openings - you have been looking at. |
||
|
i'll try My favourite opening is f4 - makes me happy and creates a bit faster games for me, or a bit more wild then e4 (now have many more games with this one as i though i might be losing too much with it and wanted to try a change) I will not share a board here but rather the options on my DB black wins 60%, white 35% not so good (and really much less than my current white wins 45% overall) now.. black reacts mostly with d5 which gives black 61% to win. the option I used most is c4 which gives black still 55% to win. followed by c6 black has 57% my d4 makes it 66% for black and so on... opponent rating keeps changing us well all the time. I am just not sure how do learn from that. and this one case if i follow for example opening with e4 (53% for black) e5 gives black 44% and so on - after 5 top moves it is white which wins with 40% (2.Bc4, Nf6 3. d3 Nc6) |
||
|
Not easy to figure the problem. Let's see if this helps: I looked at the Bird Opening. Nothing wrong with it - I used to play it to get into a Stonewall Attack which in days long past was a favorite of mine. I thought I would take it a step at a time for 5 ply, and see what showed. This looked at the GK World Database 1,d4 6.532 games, average ELO of players adopting it: 2236. Not exactly GM level, but a strong enough endorsement for you and me. Record (for every 10 games) W3 D3 L4 roughly. That's not too bad, if you like the Bird. 1... d5 The most popular response, but fewer than half of all Bird opening games (Now, THAT is interesting! You might have to count on all sorts of deviations at this point, including the From Gambit). 2,962 games, average ELO 2257 a trifle stronger opposition, but not very significant difference. Record (for every 10 games) W3 D3 L4 - the percentages hardly different from the above, really. 2.c4 ... Far from the most popular, only 20 examples in the World database. 20 games only, players adopting it, 2213 ELO Record: 8W 4D 8L Dead even! The other moves' scores aren't so encouraging. That would tempt me to look further into this line: relatively unexplored, reasonable results so far. 2...c6 But this might give me pause - the most popular reply, and seemingly very successful too 7 games only, and look at the players' ELO ratings: just 2141. Record: 2W oD 7L. Not looking so good, but we continue: 3...e3 Just 3 games with this line, by opposition at 2230 a deal stronger than those who played 2...c6. Record: 2W 0D 1L Statistically insignificanr. Those black and white bars, of course, just mark the WDL in visual form. When the numbers are small they become less significant. ( the 2-0-5 results from 7 games I wouldn't set too much store by, actually, but it is some indication. So far (if I were examining the Bird Opening with a view to playing it, I would feel inclined to try this line in actual play. For the purposes of this thread, I'll stop here, but in fact I investigated further, until it was showing but one game - a White win, I think - but without identifying an actual game. That is usually a sign that it will transpose into a game that was played with a different move order. Sure enough, after two or three 'ply' the numbers started expanding until I was looking at a position that had been played 23 times before on the database! Very positional sort of game, a kind of Anti-Stonewall position with White pawns at c4, d4, e3 and f4, well developed game. Results pretty good, but not 100%, of course. Were I exploring 1.f4 d5 2.c4, I'd find the results encouraging. But before giving it a test run in actual play, I'd run through a few games first, just to get a feel of the types and shapes the games are like to take. If I like them, they get the test runs. If I don't, I'll look at other lines. I don't know whether that quite answers the question you have in mind, but I hope it offers some help. |
||
|
|
||
|
thanks Gameknot.com I look mainly at my own and feel confused sometimes. At some points it seems i am on a really bad train, as you suggest, but if i Stick to it - most chances it will turn out to be a winning train that is where my confusion lies. also, if you are already here - could you explain me what means the Rating in my personal Database? because it cannot be made by my own rating playing those games. |
||
|
|
||
|
so how come |
||
|
Just done it - and hoping it will help instead of trawling through my history as I usually do I've never seen one of my games win lose or draw on the gameknot database Then again I've never seen anyone I know ! Do you have to retire before a game is shown !😉 Pirate Neil |
||
|
I may be wrong, but I think only games when both players are at least 1,600 elo will register. |
||
|
I suppose there are so many games on it that it maybe pot luck ! |
||
|
|
||
|
yon_cassius 01-Dec-18, 10:44 |
Presumably this just depends on what the ratings of the players were at the end of the game though; so if someone was rated 1600 at the end of a game, but had a lower rating floor than 1600, their rating could sink below 1600 at some point after the game... the game would presumably remain on the DB though. I find the personal DB one of the most useful features here. It pays to update it regularly, as this won't happen automatically. |
||
|
Personal Database |
